The Legal Asset Integrity Standard

A Framework for Fraud Prevention, Transparency, and Institutional Governance in Litigation Finance

Contents

1. 1.Executive Summary
2. 2.Part I. Legal Asset Integrity Standard: market need, integrity principles, Legal Asset Passport, LAIS, evidence confidence, hard stops, lifecycle surveillance, Atlas product architecture, fraud controls, governance, regulation, data normalization, and responsible automation.
3. 3.Part II. Quantitative Measurement Science: Legal Asset State Vector, scientific estimands, evidence science, frozen-date replay, statistical validation, survival analysis, graph analytics, risk-adjusted counterparty intelligence, portfolio digital twins, model-risk governance, and institutional framework crosswalk.
4. 4.Part III. The Institutional Asset-Management Operating System: institutional decision loop, Institutional Decision Genome, information seasoning, capital duration, staged commitment, capital routing, roll-up learning flywheel, research benchmarks, and proof of institutional operating model.
5. 5.Part IV. Capital Markets, Implementation, and Industry Leadership: financing implications, implementation roadmap, KPI framework, use cases, open-standard/proprietary-moat design, thought leadership, limitations, and conclusion.
6. 6.Appendices A-J: scorecards, event codes, hard stops, sample reports, dashboards, publication model, model validation, institutional decision record, acquisition reconstruction protocol, and quantitative definitions.

Executive Summary

Litigation finance is entering an institutionalization phase.

The market provides meaningful benefits. Funding can expand access to justice, allow plaintiffs to withstand long litigation timelines, enable law firms to finance meritorious cases, and provide investors with exposure to a differentiated asset class. At the same time, the market has attracted increasing scrutiny regarding transparency, disclosure, funder influence, data quality, conflicts, fee structures, foreign investment, and the ability of courts and counterparties to understand who has an economic interest in litigation.

The regulatory response is no longer theoretical. In June 2026, North Carolina enacted Session Law 2026-14, which prohibits the form of "litigation investment" defined in the statute for covered civil proceedings and contracts. At the federal level, proposed legislation has continued to focus on disclosure and limits on funder influence. The U.S. Government Accountability Office has repeatedly identified gaps in publicly available litigation funding data, including limited information regarding market size, rates of return, and funding arrangements.[1][2]

These developments should not be viewed only as threats to the industry. They are evidence that the industry's operating infrastructure must mature.

A credible institutional response requires more than policy statements. It requires technology, standardized data, auditable controls, and continuous evidence.

The Legal Asset Integrity Standard is built around seven propositions:

1. 1.The industry has outgrown manual diligence. Relationship underwriting and static memos cannot safely scale across thousands of legal assets, multiple originators, acquired portfolios, and institutional capital structures.
2. 2.Legal assets need verification infrastructure. Courts, liens, parties, attorneys, providers, settlement rights, payoff obligations, and economic interests must be independently mapped and reconciled.
3. 3.Fraud and integrity risk are often data-detectable. Fraud rarely arrives labeled as fraud. It appears as mismatches, unusual patterns, duplicate claims, unexplained changes, stale records, behavioral anomalies, and broken chains of evidence.
4. 4.Capital rights must remain distinct from litigation control. Financing structures require clear, provable boundaries protecting client authority, attorney independence, confidentiality, and privilege.
5. 5.Court, servicing, lien, repayment, provider, and portfolio data must be normalized. Fragmented records become institutionally useful only after identity resolution, standardization, lineage, and reconciliation.
6. 6.Investors need portfolio surveillance, not static underwriting. Legal assets evolve. Institutional risk management requires continuous monitoring, risk migration, exception management, and evidence freshness.
7. 7.The future is verified, explainable, and compliance-ready. Every material conclusion should be traceable to source evidence, model version, confidence level, decision authority, and an auditable record.

The standard introduces five operating artifacts:

• —Legal Asset Passport: the canonical record for each financed legal asset.
• —LAIS: the 0 to 100 Legal Asset Integrity Score.
• —Evidence Confidence Score: a separate measure of the quality, independence, freshness, and completeness of evidence supporting LAIS.
• —Integrity Event Ledger: a time-stamped record of changes, risk signals, verifications, approvals, exceptions, and resolutions across the asset lifecycle.
• —Integrity Exception Framework: a formal mechanism for curing, approving, escalating, or rejecting integrity deficiencies.

The standard is designed for litigation finance platforms, law firms, portfolio investors, warehouse lenders, asset-backed lenders, insurers, reinsurers, servicers, acquisition teams, boards, auditors, and regulators.

Its governing premise is simple:

Integrity must be established before economics are optimized.

The institutional extension of that premise is equally important:

Integrity, value, uncertainty, portfolio fit, and capital treatment are different measurements and must never be collapsed into one opaque score.

Accordingly, the standard separates three primary asset measures:

1. 1.LAIS Integrity Score: Can the asset be trusted, verified, governed, and monitored?
2. 2.Economic Mark: What is the probability-weighted economic value of the asset given information available at time \(t\)?
3. 3.Evidence Confidence Score: How reliable, complete, independent, and current is the evidence supporting the first two measures?

A fourth layer, Portfolio and Capital Treatment, then determines whether the asset should be acquired, sized, priced, financed, transferred, insured, syndicated, reserved, or rejected. This separation is designed to prevent high projected returns from masking weak integrity, poor evidence, or bad portfolio fit.

Part I. The Legal Asset Integrity Standard

1. Why Legal Asset Integrity Is Now an Industry Requirement

1.1 The market has crossed an institutional threshold

The early litigation finance market could operate largely through specialized judgment. Experienced operators knew attorneys, understood venues, reviewed case files, applied conservative assumptions, and funded manageable portfolios. That approach created value because legal assets are complex and expert judgment matters.

The problem is not judgment. The problem is scale.

As portfolios grow, originators consolidate, capital structures become more sophisticated, and institutional investors require deeper reporting, judgment must be translated into systems. Otherwise, the same expertise that created the market becomes a bottleneck.

Institutional markets require repeatability across four dimensions:

• —Asset identity: every financed asset can be uniquely identified and reconciled.
• —Evidence quality: material facts are supported by attributable, current, and auditable evidence.
• —Control consistency: policies are applied the same way across operators, books, jurisdictions, and vintages.
• —Portfolio visibility: risk is monitored continuously, not rediscovered during periodic reviews.

Litigation finance has reached the point where those capabilities are no longer optional operating improvements. They are prerequisites for durable scale.

1.2 Regulation is exposing the infrastructure gap

North Carolina's 2026 legislation is a significant market signal. The statute does not merely add a disclosure form or fee cap. It prohibits the defined category of litigation investment in covered circumstances and establishes enforcement remedies.[3]

At the federal level, proposed legislation in 2026 focused on public disclosure of certain third-party litigation funding and restrictions on funder influence over litigation strategy or settlement negotiations.[4]

The policy debate contains competing views. Supporters of litigation finance emphasize access to justice, risk transfer, financing flexibility, and the ability to pursue meritorious claims against better-capitalized opponents. Critics focus on opacity, cost, conflicts, influence, foreign capital, and the possibility that funding can alter litigation incentives. Both perspectives point toward the same operational conclusion:

The market needs better evidence, clearer boundaries, and stronger controls.

A mature industry should be able to demonstrate, rather than merely assert, that:

• —funded cases are authentic;
• —liens and payoff rights are known;
• —duplicate funding is detected;
• —counsel retains independent professional judgment;
• —privilege and confidentiality are protected;
• —economic rights are documented;
• —settlement flows are reconciled;
• —portfolio risk is continuously monitored;
• —investor reporting is supported by source-level evidence; and
• —model-driven decisions are explainable and auditable.

1.3 The opportunity is to lead rather than react

The strongest response to regulatory pressure is not defensive lobbying alone. It is to raise the operating standard of the market.

The industry can choose between two paths:

Reactive path: address new rules jurisdiction by jurisdiction, add disclosure forms, increase manual review, and accept rising compliance cost.

Institutional path: create a common integrity framework that continuously verifies legal assets, captures jurisdictional requirements, separates legal control from capital rights, and produces audit-ready evidence by design.

The Legal Asset Integrity Standard is designed for the second path.

2. The Industry Has Outgrown Manual Diligence

2.1 Manual diligence is a snapshot of a moving asset

A litigation asset changes after funding. Motions are filed. Judges rule. Counsel changes. Discovery progresses. Medical treatment continues. Liens grow or are negotiated. Defendants and insurers change posture. Settlement discussions emerge. Cases become inactive, dismissed, appealed, transferred, consolidated, or resolved.

A memo prepared on day one becomes less reliable every day unless the underlying evidence is refreshed.

Static underwriting therefore contains a structural weakness: it records a point-in-time belief about a dynamic asset.

2.2 The seven scaling failures of manual review

Failure 1: Source fragmentation

Relevant information is distributed across court systems, law firm records, funding platforms, lien documents, medical records, corporate databases, UCC systems, public records, emails, servicing notes, and payment systems.

Failure 2: Identity inconsistency

Names, case numbers, entities, law firms, providers, defendants, insurers, and claimants can appear in different formats across systems. Without entity resolution, duplicate exposure and contradictory records are difficult to identify.

Failure 3: Uneven operator judgment

Two experienced underwriters may interpret the same file differently. Expert judgment remains valuable, but institutions need to know which rules were applied, what evidence supported the decision, and when an exception was approved.

Failure 4: Stale information

A verified fact can become false. Counsel can withdraw. A case can settle. A lien can be filed. A corporate defendant can enter bankruptcy. A previously clean payoff chain can become disputed.

Failure 5: Portfolio blindness

Individual assets may appear acceptable while the portfolio develops dangerous concentration in one attorney, provider, venue, defendant, insurer, originator, case type, or time period.

Failure 6: Acquisition inconsistency

Roll-ups inherit different schemas, contracts, underwriting philosophies, servicing practices, and definitions. Without normalization, consolidated reporting creates false comparability.

Failure 7: Weak institutional evidence

A lender, auditor, board, regulator, or investment committee cannot rely indefinitely on "our team knows this attorney" or "the case looked strong at funding." Institutional capital requires evidence that can survive independent review.

2.3 Judgment should be encoded, not eliminated

The goal is not to automate professional judgment out of litigation finance. The goal is to preserve the best judgment, make it repeatable, detect where it is unsupported, and learn from outcomes.

A mature system should answer:

• —What did the underwriter believe?
• —What evidence supported that belief?
• —Which rules were applied?
• —Which rules were overridden?
• —Who approved the override?
• —What changed after approval?
• —Did the asset perform as expected?
• —What should the organization learn from the outcome?

That creates institutional memory rather than institutional dependence on individual memory.

3. What Is Legal Asset Integrity?

3.1 Definition

Legal Asset Integrity is the degree to which a legal asset's identity, economic rights, supporting evidence, counterparties, lifecycle status, control boundaries, and servicing record can be independently verified, continuously monitored, and governed through auditable controls.

Integrity is not the same as value.

A case can be completely authentic, fully documented, and economically unattractive. Conversely, a case can appear to have extraordinary potential value while containing integrity deficiencies that make it inappropriate to finance.

This distinction is foundational.

3.2 Four questions that must remain separate

LAIS addresses the second question. It should never be used as a substitute for legal merits analysis or return underwriting.

3.3 The institutional decision sequence

The recommended sequence is:

1. 1.Jurisdiction eligibility gate
2. 2.Identity and case verification
3. 3.Legal Asset Integrity Score
4. 4.Economic underwriting
5. 5.Portfolio fit and concentration review
6. 6.Approval and exception governance
7. 7.Continuous lifecycle surveillance
8. 8.Settlement and payoff reconciliation
9. 9.Outcome learning and model recalibration

This sequence prevents an attractive projected return from overpowering a basic integrity failure.

3.4 The Integrity-Economics Matrix

The most dangerous asset is often the one with weak integrity and a compelling projected return. The standard is explicitly designed to prevent "economics override" behavior.

4. A Taxonomy of Fraud and Integrity Failure

Fraud prevention begins with a precise taxonomy. Not every integrity failure is intentional fraud. Some failures result from poor data, operational mistakes, delayed updates, inconsistent processes, or legitimate disputes. The system should detect anomalies without prematurely labeling conduct.

The following taxonomy separates common failure modes from the signals and controls used to identify them.

4.1 Fraud is often a graph problem

Many important signals are not visible within a single case. They emerge from relationships among:

• —claimants;
• —law firms;
• —individual attorneys;
• —medical providers;
• —funding companies;
• —originators;
• —servicers;
• —defendants;
• —insurers;
• —bank accounts;
• —addresses;
• —phone numbers;
• —devices;
• —related entities; and
• —prior funded matters.

A legal asset integrity platform therefore requires both case-level analysis and graph-level analysis.

4.2 Fraud is also a time-series problem

The timing of events matters. A payment instruction changed two months before resolution is different from one changed hours before disbursement. A case with no docket activity for nine months is different from a case with a trial date and active discovery. A provider concentration that emerges gradually across a portfolio can be more meaningful than any single invoice.

The standard therefore treats chronology as evidence, not metadata.

5. The Ten Principles of the Legal Asset Integrity Standard

Principle 1: Integrity Before Economics

No expected return, settlement estimate, or strategic relationship should override an unresolved hard-stop integrity failure.

Principle 2: Independent Verification

Material facts should be verified against independent or authoritative sources wherever reasonably available. Self-reported information may be used, but its source class must remain visible.

Principle 3: Evidence Traceability

Every material score, flag, approval, and exception should be traceable to the evidence that produced it.

Principle 4: Continuous Verification

Verification is not a closing event. Assets should be rechecked as the underlying case, collateral, counsel, lien structure, and payment status evolve.

Principle 5: Separation of Capital and Legal Control

Economic rights must be structurally separated from litigation strategy, settlement authority, attorney professional judgment, and protected communications.

Principle 6: Explicit Uncertainty

A system should not create false precision. Data confidence, sparse-data conditions, missing evidence, and model uncertainty should be reported separately from the substantive score.

Principle 7: Rules Before Exceptions, Evidence Before Overrides

Policies should be encoded as rules. Exceptions should require a stated reason, supporting evidence, named authority, expiration or review date, and complete audit trail.

Principle 8: Portfolio Context

An asset that is acceptable in isolation can be unacceptable in a concentrated portfolio. Integrity monitoring must operate at asset, cohort, originator, and portfolio levels.

Principle 9: Privacy, Privilege, and Least Access

The platform should collect only data required for a defined purpose, protect privileged and sensitive information, and restrict access according to role and necessity.

Principle 10: Learn From Outcomes

Realized outcomes, losses, disputes, delays, settlements, recoveries, overrides, and false positives should feed back into rules, thresholds, models, and operating procedures.

6. The Legal Asset Passport

6.1 A canonical record for every legal asset

The Legal Asset Passport is the standardized digital identity and evidence record for a legal asset throughout its lifecycle.

It is not simply a data row. It is the authoritative record of what the organization knows, how it knows it, when it was last verified, which facts remain uncertain, what economic claims attach to the asset, and what controls have been applied.

6.2 Minimum passport schema

Identity

• —Global internal asset identifier
• —Court and docket number
• —Jurisdiction
• —Court
• —Case caption
• —Claimant / plaintiff identity token
• —Defendant identity
• —Counsel and law firm identity
• —Case type and sub-type
• —Filing date
• —Current procedural status

Economic rights

• —Funding instrument type
• —Advance amount
• —Contract date
• —Return mechanics
• —Lien or assignment status
• —Priority position
• —Payoff calculation method
• —Known competing claims
• —UCC or other filing references where relevant

Case evidence

• —Docket source
• —Key pleadings and orders
• —Relevant procedural events
• —Trial or hearing dates
• —Settlement indicators
• —Counsel attestations
• —Document fingerprints

Medical and provider data, when applicable

• —Providers
• —Treatment dates
• —Billed amounts
• —Lien amounts
• —Negotiated or expected payoff
• —Injury / treatment taxonomy
• —Provider concentration and relationship flags

Servicing

• —Last attorney update
• —Last docket verification
• —Last document refresh
• —Open exceptions
• —Payment status
• —Settlement status
• —Escalation status

Governance

• —Applicable jurisdiction rules
• —Disclosure requirements
• —Client consent indicators where applicable
• —Privilege classification
• —Control restrictions
• —Approval history
• —Human overrides

Scoring

• —Current LAIS
• —Evidence Confidence Score
• —Pillar scores
• —Economic underwriting references
• —Portfolio fit result
• —Score history
• —Active hard stops
• —Active warnings

6.3 Evidence hierarchy

Not all evidence should carry equal weight. The standard recommends classifying evidence by source quality.

Model-derived information can be valuable, but it should never be presented as an independently verified fact.

6.4 Freshness

Every material field should carry:

• —source date;
• —ingestion date;
• —last verification date;
• —expected refresh interval; and
• —stale-data threshold.

A fact can be true and still be too old for the decision being made.

7. LAIS: The Legal Asset Integrity Score

7.1 Purpose

LAIS is a standardized 0 to 100 score measuring the integrity of a legal asset.

It is designed to support:

• —pre-funding verification;
• —acquisition diligence;
• —portfolio onboarding;
• —collateral monitoring;
• —servicing prioritization;
• —fraud detection;
• —warehouse and lender reporting;
• —investor reporting;
• —governance;
• —exception management; and
• —post-outcome learning.

7.2 What LAIS does not measure

LAIS is not:

• —a legal opinion;
• —a prediction that a claimant will win;
• —an expected settlement value;
• —an IRR or MOIC forecast;
• —a substitute for attorney judgment;
• —a substitute for jurisdiction-specific legal analysis; or
• —a guarantee against fraud or loss.

A high-integrity asset can still be a poor investment. A low-integrity asset should not become fundable merely because a valuation model projects a high return.

7.3 LAIS score bands

7.4 Baseline pillar weights

Weights should be calibrated by asset class. Medical and lien integrity may be substantially more important for medical receivable or personal injury strategies, while venue and outcome strength may carry greater weight in commercial or patent litigation portfolios.

7.5 Score construction

The baseline score is a weighted combination of the eight pillar scores:

LAIS = Σ (Pillar Score × Pillar Weight)

However, raw arithmetic cannot override critical controls. LAIS operates with three additional mechanisms:

1. 1.Hard Stops: critical failures that prohibit approval or cap the score until cured.
2. 2.Evidence Confidence: a separate 0 to 100 score reflecting the quality and completeness of evidence supporting LAIS.
3. 3.Freshness Rules: automatic warnings or score constraints when material data becomes stale.

The recommended output is therefore two-dimensional:

LAIS 84 / Evidence Confidence 92

This is more honest than embedding uncertainty invisibly inside a single number.

7.6 Boundary between integrity and investment merits

Two pillars, Venue and Outcome Strength and Duration Risk, can easily become confused with investment underwriting. The standard therefore imposes a boundary.

Within LAIS, these pillars measure whether the venue, outcome, and duration assumptions are supported, stable, sufficiently evidenced, and appropriate for institutional reliance. They do not determine whether the expected recovery produces an attractive return.

Examples:

• —A venue can have weak expected economics but strong integrity because the empirical evidence is extensive, current, and internally consistent.
• —A venue can have apparently attractive projected economics but weak integrity because the relevant sample is sparse, stale, selectively reported, or poorly transferable.
• —A long-duration asset can have high integrity if the duration distribution is well evidenced and appropriately structured.
• —A short projected duration can have low integrity if it relies on resolved-only observations or unsupported assumptions.

Economic attractiveness belongs in the separate Economic Mark and return-underwriting layer. This prevents double counting and makes LAIS more defensible as an integrity standard rather than a disguised investment score.

8. LAIS Pillar Methodology

8.1 Pillar 1: Case Authenticity

Core question: Does the case exist, and do independently verifiable facts match the submitted asset file?

Subdimensions

Critical flags

• —case cannot be located in an expected authoritative source;
• —submitted docket number belongs to another matter;
• —material party mismatch;
• —matter represented as active is verified as resolved or dismissed without documented explanation;
• —document chronology is impossible or materially inconsistent.

Institutional objective

No capital should be deployed into an asset whose underlying legal matter cannot be independently identified and reconciled.

8.2 Pillar 2: Collateral Clarity

Core question: Are the financed economic rights identifiable, documented, prioritized, and free from unresolved duplicate claims?

Subdimensions

• —funding agreement completeness;
• —lien acknowledgment;
• —assignment language;
• —priority position;
• —prior funding exposure;
• —payoff formula clarity;
• —UCC or equivalent filing status where applicable;
• —conflicting claims;
• —waterfall logic;
• —enforceability review status.

Critical flags

• —unknown or disputed priority;
• —evidence of undisclosed duplicate funding;
• —missing core economic agreement;
• —material conflict between payoff records;
• —known competing claim with no resolution path.

Institutional objective

Legal finance cannot become reliable collateral if the market cannot identify who is entitled to what proceeds and in what order.

8.3 Pillar 3: Counsel Reliability

Core question: Does counsel demonstrate reliable behavior across case performance, reporting, documentation, settlement handling, and repayment obligations?

Subdimensions

• —historical funded-case performance;
• —settlement reporting timeliness;
• —response time;
• —document completeness;
• —payoff history;
• —dispute rate;
• —case duration variance;
• —concentration exposure;
• —bar / disciplinary data where lawfully and appropriately used;
• —consistency between counsel updates and external evidence.

Behavioral insight

Counsel reliability should be modeled as a longitudinal signal, not a reputation label. A firm may be strong in one venue or case type and weak in another. A new attorney may have insufficient history. The system should distinguish poor performance from sparse data.

Institutional objective

Convert anecdotal attorney reputation into evidence-based, context-sensitive operational intelligence.

8.4 Pillar 4: Medical and Lien Integrity

Core question: Where medical treatment or receivables are relevant, are treatment patterns, provider relationships, billed amounts, liens, and expected payoffs consistent with supportable recovery economics?

Subdimensions

• —treatment duration;
• —billed amount by injury / procedure type;
• —provider concentration;
• —attorney-provider relationship concentration;
• —lien amount relative to modeled recovery;
• —prior negotiated payoff behavior;
• —treatment chronology;
• —provider entity identity;
• —duplicate invoice or lien detection;
• —jurisdiction-specific lien rules.

Critical flags

• —lien amount exceeds plausible modeled recovery without explanation;
• —duplicate or conflicting medical lien claims;
• —suspicious provider / attorney / claimant clustering;
• —material treatment chronology inconsistencies;
• —unidentified lien holder or unclear payoff authority.

Institutional objective

Make medical and lien economics visible, comparable, and auditable rather than treating them as unstructured attachments to the legal case.

8.5 Pillar 5: Venue and Outcome Strength

Core question: Does the jurisdictional and procedural environment support the assumptions being used in underwriting?

Subdimensions

• —venue-level outcomes;
• —judge-specific procedural behavior where lawful and statistically supportable;
• —case-type comparables;
• —defendant characteristics;
• —insurer behavior;
• —motion patterns;
• —local procedural timing;
• —settlement distributions;
• —appeal patterns;
• —data density and sample quality.

Explainability requirement

Venue and outcome models should show cohort size, time period, comparable-case definition, calibration, and uncertainty. Sparse or biased samples must be flagged.

Institutional objective

Replace unsupported narrative valuation with transparent evidence while preserving human legal judgment.

8.6 Pillar 6: Duration Risk

Core question: Is the asset progressing within a reasonable time distribution for its procedural posture, venue, counsel, defendant, and case type?

Subdimensions

• —time since filing;
• —time since last meaningful docket event;
• —discovery status;
• —motion cadence;
• —trial setting;
• —venue-specific duration distributions;
• —counsel duration history;
• —defendant/insurer delay behavior;
• —settlement history;
• —appeal risk.

Why duration is integrity-relevant

Duration is not merely an economic variable. Material divergence between represented timing and observable case progress can indicate stale servicing, unrealistic underwriting, missing case events, or misrepresentation.

Institutional objective

Treat duration as a monitored distribution that updates with each material case event.

8.7 Pillar 7: Servicing Risk

Core question: Is the asset being monitored, documented, reconciled, and escalated with sufficient operational discipline?

Subdimensions

• —update cadence;
• —docket refresh cadence;
• —missing documents;
• —attorney communications;
• —settlement detection;
• —payoff tracking;
• —exception aging;
• —SLA compliance;
• —collateral file completeness;
• —investor reporting reconciliation.

Institutional objective

Make servicing measurable. Strong origination cannot compensate for weak lifecycle control.

8.8 Pillar 8: Control and Compliance Risk

Core question: Are economic rights appropriately structured, litigation decision-making boundaries preserved, and jurisdiction-specific governance requirements documented?

Subdimensions

• —jurisdiction eligibility;
• —funding disclosure requirements;
• —litigation-control restrictions;
• —settlement authority;
• —privilege safeguards;
• —confidentiality controls;
• —conflicts;
• —related-party relationships;
• —client consent requirements where applicable;
• —investor and lender reporting requirements;
• —approval and override history.

Critical flags

• —product prohibited in applicable jurisdiction;
• —contract appears to grant impermissible litigation control;
• —required consent or disclosure is missing;
• —material privilege or confidentiality breach;
• —unresolved conflict involving decision-makers or related parties.

Institutional objective

Prove through data and workflow that financing economics do not displace professional legal independence or required client authority.

9. Evidence Confidence Score

9.1 Why confidence must be separate

A LAIS of 85 supported almost entirely by stale self-reported information is not equivalent to a LAIS of 85 supported by current court data, signed lien acknowledgments, verified servicing records, and reconciled payment information.

The Evidence Confidence Score (ECS) measures the reliability of the evidence supporting the LAIS result.

9.2 Illustrative ECS dimensions

9.3 Confidence bands

9.4 Example

An asset might receive:

• —LAIS: 88
• —ECS: 54

The correct response is not "institutional grade." The correct response is "potentially strong integrity profile, but evidence is insufficient. Cure evidence gaps before approval."

This prevents the score from becoming a false-certainty machine.

10. Hard Stops, Warnings, and Exceptions

10.1 Hard-stop philosophy

Some conditions should not be averaged away by strong performance elsewhere.

Illustrative hard stops include:

• —legally prohibited product or structure in the applicable jurisdiction;
• —case cannot be verified;
• —known undisclosed duplicate funding;
• —unresolved ownership or lien priority conflict;
• —confirmed material misrepresentation;
• —payment instructions cannot be independently verified;
• —required consent or disclosure is absent;
• —critical identity mismatch;
• —material privilege or confidentiality control failure.

Hard stops can either:

• —block funding entirely;
• —cap LAIS below a defined threshold; or
• —require formal exception committee approval after specified cure actions.

10.2 Warnings

Warnings indicate elevated risk but do not automatically block approval. Examples:

• —stale counsel update;
• —elevated attorney concentration;
• —sparse venue data;
• —medical specials outside cohort range;
• —duration drift;
• —unresolved nonmaterial document gap;
• —unusually high human override frequency.

10.3 Exception record

Every exception should include:

• —rule triggered;
• —asset identifier;
• —reason for exception;
• —evidence supporting the exception;
• —approving authority;
• —date and time;
• —conditions imposed;
• —expiration or review date;
• —required monitoring;
• —final outcome.

The system should learn from exceptions. If the same exception occurs repeatedly, the organization either has a broken rule or a broken process.

11. Continuous Lifecycle Rescoring

A legal asset should not carry a permanent integrity score.

11.1 Lifecycle stages

Stage 1: Intake

• —verify identity;
• —determine jurisdiction eligibility;
• —collect minimum evidence;
• —identify known liens and prior funding;
• —establish initial passport.

Stage 2: Underwriting

• —calculate LAIS and ECS;
• —run economic underwriting separately;
• —evaluate portfolio fit;
• —resolve hard stops;
• —record approvals and exceptions.

Stage 3: Funding / acquisition

• —confirm final documents;
• —fingerprint contract set;
• —verify payment instructions;
• —establish final lien/payoff map;
• —create monitoring schedule.

Stage 4: Servicing

• —refresh dockets;
• —collect counsel updates;
• —monitor liens and economic claims;
• —track duration drift;
• —recalculate LAIS as evidence changes.

Stage 5: Settlement / resolution

• —detect settlement signals;
• —verify settlement status;
• —calculate payoff waterfall;
• —independently verify payment instructions;
• —reconcile receipts.

Stage 6: Post-resolution learning

• —record realized outcome;
• —compare predicted vs actual duration and recovery;
• —capture disputes and exceptions;
• —update counsel/provider histories;
• —recalibrate models and rules.

11.2 Integrity events

An Integrity Event is any new fact that can materially change the integrity profile of an asset.

Examples:

• —counsel withdrawal;
• —new lien;
• —trial date;
• —dismissal;
• —bankruptcy filing;
• —settlement notice;
• —changed bank instructions;
• —conflicting payoff request;
• —new funder discovered;
• —provider dispute;
• —missed servicing SLA;
• —jurisdiction rule change.

Material Integrity Events should trigger automatic rescoring, re-routing, or escalation.

12. The Atlas Legal Asset Integrity Platform

The standard is designed to be implementation-neutral. Atlas is one architecture capable of operationalizing it at scale.

The Atlas implementation consists of eight coordinated product modules.

12.1 Product 1: CaseVerify

Purpose

Create a verified, continuously refreshed identity and status layer for every legal asset.

Inputs

• —court and docket data;
• —filed documents;
• —party and entity records;
• —attorney and firm records;
• —user-submitted case data;
• —relevant public records;
• —case-management and servicing systems.

Core functions

• —docket validation;
• —party matching;
• —counsel-of-record verification;
• —judge and venue verification;
• —case taxonomy normalization;
• —procedural posture extraction;
• —open/closed/stayed/dismissed/appealed status;
• —related-case detection;
• —mismatch detection;
• —document fingerprinting;
• —stale-case monitoring.

Outputs

• —verified case identity;
• —Case Authenticity score;
• —mismatch report;
• —evidence lineage;
• —data freshness status;
• —active integrity warnings.

Institutional value

CaseVerify turns the most basic underwriting assumption, "this case is what the file says it is," into a provable control.

Future expansion

• —automated court-document change detection;
• —cross-jurisdiction identity resolution;
• —bankruptcy and corporate-event overlays;
• —configurable evidence requirements by asset class.

12.2 Product 2: Legal Asset Registry

Purpose

Create a permissioned record of funded legal interests, liens, assignments, and payoff claims to identify duplicate funding and priority conflicts.

Core functions

• —global asset identity;
• —funder interest registration;
• —claimant/case matching;
• —lien acknowledgment;
• —priority mapping;
• —payoff claim registration;
• —duplicate-funding detection;
• —conflict notification;
• —UCC/public-record integration where relevant;
• —immutable registration history.

Design principle

The registry should not indiscriminately expose confidential litigation information. It should be designed around minimum necessary disclosure, hashed or tokenized identifiers where appropriate, role-based access, and clear participation rules.

Institutional value

A trusted registry can reduce one of the industry's most basic structural problems: the inability to know whether another economic interest already attaches to the same expected proceeds.

Network opportunity

The registry can begin inside a single integrated platform or roll-up, where participation can be mandatory across subsidiaries. It can later expand to trusted counterparties and external participants under defined governance.

Future expansion

• —inter-funder verification protocols;
• —cryptographic proof of registration without exposing sensitive terms;
• —automated payoff coordination;
• —registry-based lender collateral verification.

12.3 Product 3: Attorney Risk Graph

Purpose

Convert attorney and law-firm behavior into longitudinal, evidence-based operational intelligence.

Inputs

• —funded case history;
• —repayment history;
• —settlement timing;
• —update responsiveness;
• —documentation quality;
• —lien disputes;
• —case outcomes;
• —duration distributions;
• —venue mix;
• —client and portfolio concentration;
• —public attorney and firm records where appropriate.

Core functions

• —attorney and firm entity resolution;
• —behavioral scoring;
• —performance segmentation by case type and venue;
• —concentration analysis;
• —settlement-reporting reliability;
• —servicing responsiveness;
• —dispute frequency;
• —anomaly detection;
• —network relationships with providers, originators, and other counterparties.

Guardrail

Attorney scoring should distinguish observed behavior from inference and should account for data sparsity. It should not become a reputational blacklist based on opaque correlations.

Institutional value

The Attorney Risk Graph converts one of the industry's largest intangible assets, relationship knowledge, into structured institutional memory.

Future expansion

• —counsel onboarding tiers;
• —dynamic servicing frequency by counsel risk;
• —attorney-level exposure limits;
• —portfolio optimization by counsel/venue pair.

12.4 Product 4: Medical Lien Integrity

Purpose

Normalize and monitor the medical economics that often sit between case value, claimant recovery, and funder repayment.

Inputs

• —provider identity;
• —treatment chronology;
• —billed amounts;
• —liens;
• —negotiated payoffs;
• —injury and treatment taxonomy;
• —provider-attorney relationships;
• —comparable settlement and recovery data;
• —jurisdictional lien rules.

Core functions

• —provider entity resolution;
• —treatment intensity analysis;
• —specials-to-recovery analysis;
• —lien-to-modeled-recovery ratios;
• —provider concentration;
• —duplicate claim detection;
• —abnormal referral pattern detection;
• —historical provider recovery analysis;
• —payoff validation.

Institutional value

This module makes a historically opaque component of personal injury economics measurable and comparable.

Future expansion

• —medical receivable valuation;
• —provider network risk maps;
• —treatment pathway benchmarking;
• —automated lien negotiation analytics;
• —insurer and settlement interaction modeling.

12.5 Product 5: SettlementLock

Purpose

Protect the point at which legal asset value becomes cash.

Why it matters

The settlement and payoff stage concentrates operational, fraud, and payment risk. A platform can underwrite a case correctly for years and still lose value if settlement is not detected, lien rights are disputed, payoff calculations are wrong, or payment instructions are compromised.

Core functions

• —settlement-event detection;
• —lien acknowledgment verification;
• —payoff calculation;
• —waterfall modeling;
• —payment instruction verification;
• —beneficiary verification;
• —dual approval workflows;
• —attorney certification;
• —escrow status;
• —dispute flags;
• —final reconciliation;
• —immutable payment-event history.

Institutional controls

High-risk payment changes should require independent out-of-band verification. No last-minute email change to bank instructions should be sufficient by itself to redirect funds.

Institutional value

SettlementLock converts payoff from a manual back-office step into a controlled asset-resolution process.

Future expansion

• —API-based bank verification;
• —smart payoff certificates;
• —automated lender reporting;
• —waterfall reconciliation across multiple economic interests.

12.6 Product 6: Portfolio Surveillance Command Center

Purpose

Provide continuous, exception-driven surveillance across every asset, cohort, originator, and portfolio.

Core metrics

• —current LAIS distribution;
• —evidence confidence distribution;
• —score migration;
• —case aging;
• —docket inactivity;
• —duration drift;
• —attorney concentration;
• —provider concentration;
• —venue concentration;
• —case-type concentration;
• —originator concentration;
• —settlement lag;
• —unresolved liens;
• —exception aging;
• —document freshness;
• —realized vs expected recovery;
• —realized vs expected duration;
• —vintage performance;
• —reserve recommendations.

Operating philosophy

The objective is not more dashboards. The objective is to surface the few exceptions that require human attention.

The command center should answer:

• —What changed?
• —Why does it matter?
• —Which assets need intervention?
• —Which portfolio assumptions are drifting?
• —What action is required?
• —Who owns that action?

Institutional value

This is the layer that allows warehouse lenders, investment committees, boards, and portfolio managers to view legal assets as continuously monitored collateral rather than periodic spreadsheets.

12.7 Product 7: Disclosure and Governance Module

Purpose

Translate jurisdictional rules, contractual restrictions, governance policies, and disclosure obligations into repeatable workflows.

Core functions

• —jurisdiction eligibility engine;
• —prohibited-structure rules;
• —disclosure requirement mapping;
• —consent workflow;
• —litigation-control certification;
• —privilege classification;
• —conflict checks;
• —related-party logs;
• —investor and lender reporting rules;
• —approval matrices;
• —audit-ready disclosure packet generation.

Jurisdiction Rules Engine

The rules engine should treat law and regulation as versioned operating logic. A rule change should be capable of identifying:

• —which products are affected;
• —which jurisdictions are affected;
• —which existing assets require review;
• —which new transactions must be blocked or modified;
• —which disclosures or approvals become mandatory.

Institutional value

This module transforms compliance from a manual legal review at the edge of the process into embedded control infrastructure.

12.8 Product 8: Model Governance and Explainability Layer

Purpose

Ensure every model-driven decision can be understood, reproduced, challenged, monitored, and governed.

Core functions

• —model registry;
• —model versioning;
• —feature lineage;
• —training-data provenance;
• —validation history;
• —performance monitoring;
• —calibration tracking;
• —drift detection;
• —sparse-data warnings;
• —comparable-cohort display;
• —confidence intervals;
• —human override records;
• —rollback capability.

Explainability packet

Every material model output should be capable of generating a compact explanation showing:

• —model name and version;
• —scoring date;
• —data used;
• —unavailable data;
• —top contributing factors;
• —relevant comparables;
• —confidence / uncertainty;
• —applicable hard stops;
• —human decision and any override.

Institutional value

This is the control that separates model-driven decisioning from black-box automation.

13. The Fraud Prevention Operating Model

Technology alone does not prevent fraud. Strong systems combine controls, accountability, and evidence.

13.1 Preventive controls

Preventive controls seek to stop bad assets or bad actions before capital is deployed.

Examples:

• —jurisdiction eligibility gate;
• —case verification;
• —identity verification;
• —duplicate-funding registry check;
• —lien acknowledgment;
• —required document set;
• —attorney onboarding tier;
• —role-based access;
• —separation of duties;
• —payment beneficiary verification.

13.2 Detective controls

Detective controls identify changes or anomalies after an asset enters the portfolio.

Examples:

• —continuous docket monitoring;
• —settlement signal detection;
• —duration drift;
• —provider anomalies;
• —counsel update mismatch;
• —concentration changes;
• —new lien detection;
• —bank instruction changes;
• —model drift;
• —unusual override patterns.

13.3 Corrective controls

Corrective controls determine how the organization responds.

Examples:

• —freeze additional funding;
• —enhanced diligence;
• —contact counsel;
• —obtain updated lien acknowledgment;
• —resize exposure;
• —increase reserve;
• —escalate to integrity committee;
• —suspend originator or provider;
• —legal/compliance review;
• —recover, reconcile, or pursue contractual remedies.

13.4 Three lines of defense

A mature platform should maintain independent roles:

1. 1.First line: origination, underwriting, servicing, and business operations own day-to-day controls.
2. 2.Second line: risk, compliance, model governance, and integrity teams define rules and challenge exceptions.
3. 3.Third line: internal audit or independent review tests whether controls operate as designed.

The specific organization can vary by scale, but the principle should remain: the same person should not originate, approve, override, and certify the same exposure without independent challenge.

14. Portfolio-Level Integrity

14.1 Integrity is not only an asset characteristic

Portfolio risk emerges from common dependencies. Even individually verified assets can create systemic exposure when concentrated around the same law firm, provider, defendant, insurer, venue, funding source, servicer, or data dependency.

14.2 Required concentration views

The standard recommends monitoring concentration by:

• —attorney;
• —law firm;
• —medical provider;
• —originator;
• —servicer;
• —jurisdiction;
• —county / venue;
• —judge where statistically meaningful;
• —case type;
• —defendant;
• —insurer;
• —injury / treatment category;
• —vintage;
• —duration bucket;
• —LAIS band;
• —evidence confidence band;
• —unresolved exception type.

14.3 Risk migration

A portfolio report should show not only current scores but movement:

• —LAIS upgrades;
• —LAIS downgrades;
• —new hard stops;
• —cured exceptions;
• —duration migration;
• —evidence-confidence deterioration;
• —new concentration breaches;
• —realized-loss clusters.

Risk migration is often more informative than a static portfolio average.

14.4 Acquisition and roll-up diligence

The standard is particularly valuable when acquiring funding books or operating companies.

A standardized onboarding process can:

1. 1.ingest the acquired book;
2. 2.resolve case and party identities;
3. 3.reconstruct collateral and payoff rights;
4. 4.verify dockets;
5. 5.score counsel and providers;
6. 6.calculate LAIS and ECS;
7. 7.identify duplicate exposures;
8. 8.quantify missing data;
9. 9.segment assets by remediation priority;
10. 10.map legacy servicing to a common operating standard.

This turns acquisition diligence from sample-based review into portfolio-level evidence reconstruction.

15. Governance and Separation of Litigation Control

15.1 Capital is not counsel

A funder may underwrite economic risk, monitor its contractual position, and receive permitted information without becoming the decision-maker in litigation.

American Bar Association materials continue to emphasize attorney independence, client authority, privilege, confidentiality, and caution around funder influence.[5]

The standard therefore requires explicit documentation of decision rights.

15.2 Decision-rights matrix

15.3 Information boundary

The platform should distinguish:

• —public case data;
• —confidential but nonprivileged information;
• —privileged information;
• —personal and medical information;
• —investor-confidential information;
• —model-proprietary information.

Access should be purpose-limited and logged.

15.4 Governance evidence

A platform should be able to demonstrate:

• —who accessed information;
• —why access was permitted;
• —what decision was made;
• —what rule governed the decision;
• —whether counsel/client authority was preserved;
• —what disclosures were made;
• —what exceptions were approved.

This is what makes the control boundary provable.

16. Regulatory and Disclosure Readiness

16.1 A jurisdiction-specific market requires a jurisdiction-specific control layer

The United States does not operate under one uniform litigation finance regime. Rules can differ across states, federal courts, judges, case types, consumer/commercial products, professional conduct rules, disclosure obligations, champerty doctrines, usury frameworks, and specific funding statutes.

A scalable platform should never treat compliance as a single national yes/no flag.

16.2 The jurisdiction eligibility gate

Before an asset enters underwriting, the system should determine:

• —governing jurisdiction;
• —product type;
• —consumer vs commercial classification;
• —prohibited structures;
• —required disclosures;
• —fee / return restrictions where applicable;
• —consent requirements;
• —control restrictions;
• —discovery / court disclosure obligations;
• —UCC / lien requirements;
• —special mass tort or class action requirements;
• —effective date of applicable rules.

16.3 North Carolina as a signal

North Carolina Session Law 2026-14 illustrates why this capability matters. A legal change can convert a previously plausible product into a prohibited structure for covered transactions.[3]

An institutional system should be capable of translating that change into operating action:

• —block affected new transactions;
• —identify contracts potentially affected by amendment or renewal;
• —notify legal/compliance owners;
• —preserve evidence of the rule version applied;
• —update product availability and workflow automatically.

16.4 Disclosure packet

Where disclosure is required, the system should be capable of generating a privilege-sensitive packet containing only what is necessary, such as:

• —existence of funding;
• —funder identity where required;
• —economic interest summary where required;
• —certification of control boundaries;
• —required ownership or foreign-interest information;
• —applicable client consent;
• —versioned legal basis for disclosure.

The objective is controlled transparency, not indiscriminate exposure of litigation strategy or privileged material.

17. Data Normalization and Interoperability

17.1 Why normalization is strategic

Legal finance data is heterogeneous by nature. The platform cannot wait for the market to standardize itself.

The integrity layer must create common representations for:

• —courts;
• —dockets;
• —cases;
• —parties;
• —attorneys;
• —law firms;
• —judges;
• —providers;
• —defendants;
• —insurers;
• —liens;
• —funding agreements;
• —payment events;
• —procedural events;
• —settlements;
• —recoveries;
• —servicing activities;
• —exceptions;
• —model outputs.

17.2 Entity resolution

A robust system should assign persistent identifiers to entities even when names change or differ across sources.

Examples:

• —"ABC Law LLP" vs "ABC Law, L.L.P.";
• —claimant name variants;
• —provider DBA names;
• —corporate subsidiaries;
• —insurer affiliates;
• —attorney changes across firms.

Entity resolution is a prerequisite for detecting concentration, conflicts, duplication, and historical behavior.

17.3 Source lineage

Every normalized field should preserve a link to its origin. Normalization must never destroy provenance.

The platform should know:

• —original value;
• —normalized value;
• —source;
• —transformation applied;
• —date received;
• —confidence;
• —reviewer where applicable.

17.4 Integration strategy

The architecture should support authoritative court, legal research, corporate, UCC, public-record, identity, banking, and payment providers through APIs and structured ingestion.

Commercial providers may include legal and compliance data platforms, court-record services, filing providers, and financial data networks. Public references to specific commercial partnerships should be made only after applicable agreements and permissions are finalized.

This allows the standard to remain provider-neutral while Atlas maintains a proprietary integration and intelligence advantage.

18. Model Governance, Explainability, and Responsible Automation

18.1 The model must never outrun the evidence

Legal finance models operate in a high-context environment. Sparse outcomes, changing law, selection effects, sealed matters, inconsistent settlement reporting, and jurisdictional differences can create false confidence.

The standard requires explicit model governance.

18.2 Minimum governance requirements

Every production model should have:

• —documented purpose;
• —owner;
• —approved use cases;
• —prohibited use cases;
• —training-data lineage;
• —validation dataset;
• —performance metrics;
• —calibration history;
• —deployment date;
• —version history;
• —monitoring thresholds;
• —fallback procedure;
• —retirement criteria.

18.3 Model drift

Models should be monitored for:

• —input drift;
• —population drift;
• —outcome drift;
• —calibration deterioration;
• —jurisdiction-specific degradation;
• —changing data coverage;
• —label leakage;
• —override frequency.

18.4 Human overrides

Human judgment remains essential, but overrides must be visible.

The platform should measure:

• —who overrides;
• —what is overridden;
• —why;
• —how often;
• —whether overrides outperform or underperform the model;
• —whether certain counterparties receive systematically favorable overrides.

Override analysis is both a model-improvement tool and a fraud/control tool.

Part II. Quantitative Measurement Science

19. The Institutional Measurement Architecture

19.1 A legal asset is a dynamic state, not a static file

A legal asset should be represented as a time-indexed state rather than a folder of documents or a one-time underwriting memo. The minimum institutional representation is a Legal Asset State Vector:

\[ X_t = \{A_t, L_t, C_t, M_t, V_t, D_t, S_t, G_t, E_t\} \]

where:

• —\(A_t\) = case authenticity and procedural state;
• —\(L_t\) = collateral, lien, priority, and payoff-right state;
• —\(C_t\) = counsel and law-firm reliability state;
• —\(M_t\) = medical, provider, and lien-integrity state where applicable;
• —\(V_t\) = venue, judge, defendant, insurer, and outcome environment;
• —\(D_t\) = duration and time-to-resolution state;
• —\(S_t\) = servicing and documentation state;
• —\(G_t\) = governance, control, jurisdiction, and compliance state; and
• —\(E_t\) = evidence quality, provenance, completeness, and freshness state.

Each element must be time stamped. Every material state variable should identify its source, confidence, lineage, last verification date, model dependency, and whether it was independently corroborated.

The purpose is not mathematical decoration. The state-vector model changes how institutions operate. It creates one canonical representation from which underwriting, servicing, surveillance, valuation, portfolio construction, financing, and reporting can draw.

19.2 Three measurements must remain separate

Institutional legal asset management should never collapse integrity, value, and confidence into a single score.

Measure 1: Legal Asset Integrity Score

\[ LAIS_t \in [0,100] \]

LAIS estimates the integrity condition of the asset at time \(t\). It answers whether the asset is sufficiently authentic, documented, governable, and monitorable for institutional treatment.

Measure 2: Economic Mark

\[ Mark_t = E\left[\frac{NCF_T}{(1+r)^T}\mid\mathcal{F}_t\right] \]

where \(NCF_T\) is the net realizable cash flow at resolution, \(T\) is the stochastic time to realization, \(r\) is the applicable discount or required return framework, and \(\mathcal{F}_t\) is the information set available at time \(t\).

The mark should normally be expressed as a distribution rather than a single-point forecast. A minimum institutional output is P10, P50, and P90, together with the expected value and the assumptions driving tail behavior.

Measure 3: Evidence Confidence Score

\[ ECS_t \in [0,100] \]

ECS measures whether the information supporting LAIS and the economic mark is authoritative, complete, independent, fresh, mutually consistent, and relevant to the decision being made.

An asset can therefore have high integrity and low confidence, or low integrity and high confidence. The latter is particularly important because strong evidence can establish that an asset is weak. Confidence is not a measure of attractiveness.

19.3 Institutional asset state card

The standard recommends that every asset be summarized using a compact state card:

LAIS 92 | ECS 96 | Mark $84,000 | P10 $55,000 | P50 $81,000 | P90 $129,000 | Median Duration 18.4 months | Integrity Trend Stable | Portfolio Fit Acceptable

The purpose of the state card is to force separation between what is known, what is believed, how uncertain the belief is, how the asset is changing, and whether the exposure fits the portfolio.

19.4 No score may bypass a hard stop

Quantitative scoring must never turn a categorical integrity failure into an averaged number. An asset with a falsified identity, prohibited jurisdictional structure, unresolved priority conflict, or impermissible litigation-control provision cannot become acceptable merely because other pillars score highly.

Formally:

\[ Eligibility_t = 0 \Rightarrow Decision_t \neq Fund \]

regardless of \(LAIS_t\), expected return, or portfolio diversification benefit.

This is a fundamental institutional control against model-induced rationalization.

20. Defining a Scientific LAIS Estimand

20.1 The score must ultimately predict something observable

A durable scoring standard needs more than expert weights. It needs an empirical target against which the score can be tested.

The recommended long-term estimand is the probability of Material Integrity Failure, or MIF, within a defined horizon.

\[ p^{MIF}_t=P(MIF_{t+h}=1\mid X_t) \]

where \(h\) is the monitoring horizon and MIF represents an event severe enough to materially impair ownership, enforceability, collectability, servicing, governance, or confidence in the financed asset.

LAIS should then be a transparent, monotonic transformation of calibrated MIF risk plus governed non-compensatory controls. The public methodology should not imply that a LAIS of 90 mechanically equals a 10% failure probability unless empirical calibration supports that interpretation. This preserves the usability of a 0 to 100 score while keeping the underlying scientific target explicit.

20.2 Material Integrity Failure taxonomy

A MIF should be precisely defined through event codes. Examples include:

• —verified case identity materially inconsistent with submitted representation;
• —undisclosed senior or competing economic interest;
• —confirmed duplicate funding or undisclosed encumbrance;
• —settlement diversion or unauthorized payoff redirection;
• —material servicing breakdown that prevents reliable asset monitoring;
• —material misrepresentation by a counterparty;
• —material provider or lien anomaly that impairs expected collectability;
• —impermissible funder control or contract provision;
• —jurisdictional prohibition or material unenforceability event;
• —document falsification or chain-of-title defect;
• —material repayment failure inconsistent with represented settlement proceeds; or
• —other board-approved integrity event meeting a defined materiality threshold.

The taxonomy should distinguish fraud, error, operational failure, legal dispute, data deficiency, and policy violation. Statistical anomaly should never be treated as proof of misconduct.

20.3 Calibration objective

Once sufficient outcomes exist, LAIS bands should have measurable empirical meaning.

For example, if a group of assets enters the 90 to 94 band, observed MIF frequency over the chosen horizon should converge toward the band’s modeled expectation. Calibration should be assessed by vintage, asset class, originator, jurisdiction, and material subpopulation.

The standard should publish calibration statistics without revealing proprietary model features.

20.4 Score stability and responsiveness

A useful integrity score must balance two competing objectives:

• —stability, so minor data noise does not create unnecessary score volatility; and
• —responsiveness, so material new evidence produces immediate risk migration.

Every change in score should therefore be decomposable into:

\[ \Delta LAIS_t = \Delta Evidence + \Delta AssetState + \Delta Model + \Delta Policy \]

This allows an investment committee or auditor to understand whether the asset changed, the evidence changed, the model changed, or the institution changed its policy.

20.5 Weighting methodology

The published LAIS weights should initially be treated as a transparent baseline, not permanent truth. Over time, weights can be refined using observed failure data subject to governance constraints.

Recommended hierarchy:

1. 1.expert-informed initial weights;
2. 2.empirical monotonicity testing;
3. 3.out-of-time validation;
4. 4.sensitivity analysis;
5. 5.challenger specifications;
6. 6.governance approval before production use; and
7. 7.periodic recalibration only where evidence supports a change.

The public standard may specify pillar definitions and minimum controls while proprietary implementations retain detailed feature engineering, model coefficients, anomaly logic, and ensemble construction.

21. Evidence Science and Data Quality

21.1 Evidence is an input to risk, not administrative metadata

Legal finance frequently treats documents as binary: present or missing. Institutional analysis should treat evidence as a measurable object.

Each material fact should be characterized by:

• —authority of source;
• —independence from the party benefiting from the assertion;
• —freshness;
• —completeness;
• —internal consistency;
• —corroboration by unrelated sources;
• —directness versus inference;
• —jurisdictional relevance;
• —machine readability;
• —chain of custody; and
• —privilege or access restrictions.

21.2 Evidence Confidence Score construction

An illustrative formulation is:

\[ ECS_t = 100\times( w_aA + w_iI + w_fF + w_cC + w_xX + w_pP ) \]

where:

• —\(A\) = authority;
• —\(I\) = independence;
• —\(F\) = freshness;
• —\(C\) = completeness;
• —\(X\) = cross-source consistency; and
• —\(P\) = provenance quality.

Weights should vary by evidence type and intended use. A court docket may have high authority for procedural status but low authority for medical treatment detail. Evidence quality is contextual.

21.3 Freshness decay

For facts that can change, evidence confidence should decay as the evidence ages.

An illustrative decay function is:

\[ F(t)=e^{-\lambda \Delta t} \]

where \(\lambda\) depends on the volatility of the field. Counsel-of-record status may require more frequent reverification than a historical filing date. Payoff instructions may require near-real-time confirmation at settlement.

21.4 Contradiction as a first-class signal

A sophisticated system should not simply choose one data source when two sources disagree. It should create a contradiction event.

Examples:

• —attorney-submitted case status differs from the court record;
• —servicing balance differs from contractual accrual terms;
• —lien amount differs across payoff documents;
• —law firm record indicates settlement while the servicer still classifies the asset as active;
• —payment destination changes after settlement notice;
• —provider data materially diverges from historical pattern for similar injuries.

Contradiction density should itself be measurable and can become an integrity predictor.

21.5 Alignment with institutional risk-data principles

BCBS 239 identifies governance, data architecture, accuracy and integrity, completeness, timeliness, adaptability, comprehensiveness, clarity, frequency, and distribution as core principles of effective risk data aggregation and reporting. LAIS applies similar discipline to legal assets, without implying that litigation finance platforms are banks or subject to banking standards.

The practical lesson is universal: fragmented data becomes institutional risk information only when it can be aggregated accurately, quickly, consistently, and across organizational boundaries.

22. Temporal Validation and the Replay Laboratory

22.1 The fundamental rule: no future information

Legal asset models are highly vulnerable to leakage because future case events are often embedded in current databases. A model can appear extraordinarily accurate if it is inadvertently allowed to see information that was not available at the historical decision date.

Every institutional validation should therefore enforce a decision-time information boundary.

For a historical decision made at time \(t_0\), the model may use only evidence that existed and was reasonably available at or before \(t_0\).

22.2 Frozen-date replay

The recommended validation method is frozen-date replay.

For each historical asset:

1. 1.identify the original decision date;
2. 2.reconstruct the evidence set available on that date;
3. 3.freeze external and internal data to the historical state;
4. 4.run the historical version or a clearly labeled challenger model;
5. 5.record LAIS, ECS, mark, duration distribution, and recommended decision;
6. 6.compare with the actual institutional decision; and
7. 7.compare both with realized outcomes.

This produces four views:

22.3 Why replay matters

Replay answers questions that ordinary backtests cannot:

• —Did the model identify risks that humans missed?
• —Did experienced operators see information not represented in the model?
• —Did a human override add value or destroy value?
• —Which data sources materially improved decision quality?
• —Which model features looked predictive only because of leakage?
• —Did the executed structure compensate for asset weakness?
• —Did better servicing improve an otherwise average asset?

The goal is not to prove that models beat people. The goal is to determine where machines, rules, and expert judgment each add incremental value.

22.4 Required validation splits

At minimum, validation should include:

• —out-of-time holdout;
• —out-of-originator holdout where practical;
• —jurisdictional holdout or transfer testing;
• —vintage analysis;
• —acquisition-book transfer testing;
• —low-volume subgroup analysis; and
• —stress-period performance.

Random train-test splits alone are insufficient for institutional claims about future performance.

23. Statistical Performance and Model Validation

23.1 Measure discrimination and calibration separately

A model can rank assets correctly but assign bad probabilities. It can also produce well-calibrated averages while failing to distinguish strong from weak assets. Institutional validation requires both.

For classification and integrity-event models, recommended metrics include:

• —ROC-AUC;
• —precision-recall AUC;
• —recall at defined review capacity;
• —precision at top-risk percentile;
• —Brier score;
• —calibration intercept and slope;
• —expected calibration error;
• —false-negative rate; and
• —cost-weighted error.

23.2 Error cost is asymmetric

Fraud and integrity detection are rare-event problems. Overall accuracy can be meaningless.

Missing a duplicate lien or settlement diversion can be materially more expensive than escalating several clean assets for manual review.

The operating threshold should therefore minimize expected economic loss:

\[ Threshold^* = \arg\min_{\tau}\left[C_{FN}FN(\tau)+C_{FP}FP(\tau)+C_RReview(\tau)\right] \]

where \(C_{FN}\) is the cost of a false negative, \(C_{FP}\) is the cost of a false positive, and \(C_R\) is the cost of human review.

23.3 Monetary prediction metrics

For recovery and asset-value models, evaluate:

• —mean absolute error;
• —median absolute error;
• —weighted absolute percentage error where denominators are stable;
• —calibration of P10/P50/P90 intervals;
• —expected-value bias;
• —tail-loss underprediction;
• —error by venue, originator, counsel, case type, and vintage; and
• —economic loss from mispricing.

A model should not be accepted because average error is low if its largest errors cluster in precisely the exposures that drive portfolio losses.

23.4 Model uncertainty

Every material prediction should distinguish, where practical:

• —aleatoric uncertainty, arising from irreducible randomness in litigation outcomes; and
• —epistemic uncertainty, arising from limited data, model uncertainty, sparse jurisdictions, new asset types, or missing evidence.

High epistemic uncertainty should lead to lower confidence, smaller sizing, wider reserves, additional diligence, or human escalation rather than false precision.

23.5 Champion-challenger discipline

Material models should have documented challenger approaches. A challenger can be simpler than the production model. In fact, simple benchmark models are valuable because they reveal whether complexity creates real incremental value.

A model should be able to answer:

What does this model do better than a simple, transparent benchmark after accounting for complexity and operational risk?

24. Survival, Duration, and Competing-Risk Science

24.1 Duration is not a completed-case average

Legal asset datasets are naturally censored. At any observation date, many cases remain unresolved. Training only on resolved matters systematically favors faster cases and can materially understate duration.

The correct object is a survival function:

\[ S(t|X)=P(T>t|X) \]

where \(T\) is time to the relevant legal or economic event.

24.2 Minimum duration toolkit

Depending on the asset class and data volume, institutional duration analysis may include:

• —Kaplan-Meier estimators;
• —Cox proportional-hazards models;
• —accelerated failure-time models;
• —competing-risk models;
• —random survival forests;
• —gradient-based survival models; and
• —Bayesian survival models.

No single method should be presumed superior. Model selection should follow empirical validation, interpretability needs, and use case.

24.3 Competing risks

Cases can leave the active state through different pathways:

• —settlement;
• —trial judgment;
• —dismissal;
• —withdrawal;
• —adverse procedural event;
• —bankruptcy-related interruption;
• —transfer or consolidation; or
• —other termination states.

Treating every exit as the same event can distort economic duration and loss estimates. Where data support it, Atlas should model cause-specific hazards or cumulative incidence functions.

24.4 Duration as an economic and financing variable

Duration affects:

• —IRR;
• —liquidity usage;
• —facility borrowing base;
• —advance rates;
• —reserve needs;
• —expected loss;
• —portfolio concentration;
• —capital recycling; and
• —value of optionality.

The standard therefore treats duration as both a case variable and a capital-management variable.

25. Graph Analytics, Entity Resolution, and Fraud Science

25.1 Integrity failures occur across relationships

Many fraud and integrity risks are invisible when an asset is viewed in isolation. A single attorney-provider relationship may be legitimate. The same relationship repeated across hundreds of unusually structured cases can be informative.

A Legal Asset Knowledge Graph should represent entities such as:

• —claimant;
• —case;
• —attorney;
• —law firm;
• —medical provider;
• —lienholder;
• —litigation funder;
• —originator;
• —servicer;
• —defendant;
• —insurer;
• —court;
• —judge;
• —payment account or destination token; and
• —affiliated business entity.

Edges encode observed relationships, contracts, referrals, representation, payments, liens, co-occurrence, and historical performance.

25.2 Network-level signals

Graph analytics can identify:

• —possible duplicate funding across entities;
• —hidden concentration that ordinary portfolio reports miss;
• —unusual attorney-provider clusters;
• —repeated payment destinations across unrelated matters;
• —abrupt changes in network behavior;
• —abnormal referral loops;
• —concentration in weak servicers;
• —correlated settlement delays;
• —shared lien counterparties with repeated disputes; and
• —suspiciously similar documentation patterns.

25.3 Entity resolution must carry confidence

Identity resolution should not silently merge records. Each match should have a confidence level and evidence basis.

For example:

\[ P(Entity_i = Entity_j | name, address, docket, firm, phone, identifiers, relationships) \]

Low-confidence matches should remain probabilistic or enter review. False merges can create as much risk as missed matches.

25.4 Anomaly is not accusation

The system should use language such as:

• —anomaly;
• —outlier;
• —unexplained relationship;
• —inconsistent pattern;
• —elevated review signal; or
• —integrity exception.

Fraud or misconduct labels should require separate evidentiary and governance thresholds.

This distinction is essential to responsible institutional use.

26. Risk-Adjusted Counsel and Provider Intelligence

26.1 Raw outcomes are not performance

Counsel who handle stronger cases, more favorable venues, or better-insured defendants may have higher realized recoveries without generating superior performance.

Institutional scoring should therefore compare observed outcomes with risk-adjusted expected outcomes.

An illustrative hierarchical model is:

\[ Y_{ij}=\alpha + \beta X_{ij}+u_j+\epsilon_{ij} \]

where \(Y_{ij}\) is the outcome for case \(i\) handled by counsel or provider \(j\), \(X_{ij}\) represents case-level characteristics, and \(u_j\) estimates the entity-specific contribution after adjusting for case mix.

26.2 Partial pooling

Low-volume attorneys and providers should not receive extreme ratings from a handful of observations. Hierarchical models allow estimates for thin-data entities to shrink toward population means until sufficient evidence accumulates.

This is both statistically sound and operationally fairer.

26.3 Counsel reliability should be multidimensional

Risk-adjusted counsel intelligence may include:

• —settlement or recovery performance relative to expected;
• —duration relative to expected;
• —servicing responsiveness;
• —documentation completeness;
• —lien-dispute incidence;
• —settlement-reporting timeliness;
• —integrity-event history;
• —concentration exposure; and
• —stability across vintages.

The goal is not a public reputation score. It is an institutional counterparty-risk measure used for underwriting, monitoring, concentration, and diligence.

27. Portfolio Science and Legal Asset Digital Twins

27.1 A good asset can be a bad portfolio addition

Institutional asset management requires marginal analysis.

For a candidate asset \(a\) and existing portfolio \(P\):

\[ \Delta Risk_a = Risk(P \cup a)-Risk(P) \]

The investment decision should consider not only standalone expected return but also the change in concentration, duration, liquidity, downside, and correlation.

27.2 Portfolio state variables

At minimum, portfolio analysis should measure exposure by:

• —law firm and attorney;
• —originator;
• —servicer;
• —medical provider;
• —lienholder;
• —venue and jurisdiction;
• —defendant and insurer;
• —case type;
• —vintage;
• —duration band;
• —LAIS band;
• —ECS band;
• —facility eligibility;
• —status and procedural posture; and
• —expected cash-flow period.

27.3 Correlation is often latent

Legal assets are sometimes described as uncorrelated. At the macro level, that can be directionally attractive, but individual legal portfolios can contain substantial hidden correlation.

Examples include:

• —one law firm controlling a large share of cases;
• —one medical provider affecting lien economics across many claims;
• —one insurer influencing settlement timing;
• —one venue changing procedural practice;
• —one acquisition source generating common underwriting bias;
• —one servicer producing common data-quality failure; or
• —one funding structure creating common refinancing risk.

The graph and portfolio layers should therefore be integrated.

27.4 Monte Carlo portfolio simulation

A portfolio digital twin should simulate thousands of plausible futures for:

• —recovery amount;
• —timing;
• —loss;
• —settlement pathway;
• —lien recovery;
• —servicing interruption;
• —refinancing conditions;
• —correlation shocks; and
• —regulatory or jurisdictional events.

Outputs can include:

• —P5/P10/P50/P90/P95 cash collections;
• —expected IRR and MOIC distributions;
• —expected shortfall;
• —liquidity-at-risk;
• —facility utilization;
• —borrowing-base sensitivity;
• —capital-at-risk;
• —duration distribution; and
• —concentration under stress.

27.5 Stress tests

In addition to stochastic simulation, portfolios should be subjected to deterministic stresses such as:

• —25% lower settlements;
• —6, 12, and 18 months of duration extension;
• —simultaneous degradation of top counsel concentration;
• —medical-lien recovery shock;
• —servicing interruption;
• —warehouse advance-rate reduction;
• —loss of facility eligibility for selected assets;
• —adverse venue rule change; and
• —delayed capital-markets exit.

The objective is to understand how quickly an apparently attractive portfolio can become a liquidity problem.

28. Institutional Model-Risk Governance

28.1 Apply financial-institution discipline without pretending to be a bank

The Federal Reserve, OCC, and FDIC issued revised interagency model-risk management guidance in April 2026, emphasizing a risk-based approach to model development, use, validation, monitoring, governance, controls, and third-party models. NIST's AI Risk Management Framework and ISO/IEC 42001 provide complementary principles for trustworthy and governed AI systems.

LAIS should borrow the discipline of these frameworks without claiming regulatory equivalence or certification where none exists.

28.2 Model inventory

Every production model should have a unique identifier and inventory record containing:

• —owner;
• —business purpose;
• —model tier;
• —methodology;
• —intended users;
• —decision materiality;
• —inputs and dependencies;
• —training window;
• —validation status;
• —production version;
• —known limitations;
• —monitoring thresholds;
• —prohibited uses;
• —third-party dependencies; and
• —retirement criteria.

28.3 Model materiality tiers

Tier 1: Capital and eligibility models

Models that can materially affect funding, acquisition price, capital allocation, facility eligibility, reserves, or portfolio construction.

Controls should include independent validation, approval before deployment, documented challenger analysis, ongoing monitoring, and event-driven revalidation.

Tier 2: Surveillance and prioritization models

Models that prioritize review, detect anomalies, or recommend interventions but do not independently authorize capital deployment.

Controls should include testing, drift monitoring, documented thresholds, and human escalation.

Tier 3: Workflow and productivity models

Models that summarize, classify, route, extract, or assist with low-materiality workflow.

Controls may be proportionate but should still include access, privacy, lineage, and quality requirements.

28.4 Independent validation

Validation should be organizationally independent from model development to the extent practical and should assess:

• —conceptual soundness;
• —data appropriateness;
• —implementation correctness;
• —outcome performance;
• —robustness;
• —limitations;
• —sensitivity;
• —stability;
• —bias and subgroup behavior where relevant;
• —override patterns; and
• —whether the model remains fit for intended purpose.

28.5 Model change governance

No material production model should change invisibly.

The model ledger should record:

\[ Version \rightarrow Change \rightarrow Rationale \rightarrow Validation \rightarrow Approval \rightarrow EffectiveDate \]

Every historical decision should remain reproducible under the model version used at the time.

28.6 Drift and regime detection

Atlas should monitor both data drift and performance drift across:

• —jurisdictions;
• —judges;
• —attorneys;
• —providers;
• —insurers;
• —case types;
• —originators;
• —servicing channels; and
• —vintages.

Potential methods include population-stability metrics, distribution-distance measures, calibration drift, residual analysis, change-point detection, and performance degradation alerts.

A drift alert does not automatically invalidate a model. It triggers investigation, challenger testing, or temporary limits where material.

28.7 Human override governance

Human judgment is a controlled model input, not an escape hatch.

Overrides should require:

• —reason code;
• —narrative rationale;
• —decision authority;
• —evidence cited;
• —magnitude of change;
• —expiration or review date where applicable; and
• —realized-outcome tracking.

Over time the organization should measure whether overrides systematically add or subtract value.

29. Institutional Crosswalk: Legal Asset Integrity and Established Risk Frameworks

The Legal Asset Integrity Standard is purpose-built for legal assets. It is not a banking regulation, accounting standard, rating methodology, or AI certification. Nevertheless, its design should be legible to institutional stakeholders accustomed to established governance frameworks.

The purpose of the crosswalk is not regulatory arbitrage. It is to ensure that a lender, board member, institutional allocator, auditor, or risk committee can recognize familiar control concepts when evaluating a new asset class.

Part III. The Institutional Asset-Management Operating System

30. From Legal Asset Integrity to Institutional Asset Management

30.1 LAIS is the trust layer, not the whole operating system

A Legal Asset Integrity Standard solves a foundational problem: whether an asset can be trusted and governed. An institutional asset manager must answer several additional questions:

• —What is the asset worth now?
• —How uncertain is that mark?
• —How should the exposure be structured?
• —Does it fit the portfolio?
• —Which capital source should own or finance it?
• —What new evidence would change the decision?
• —When can capital be refinanced or recycled?
• —What did the organization learn from the outcome?

An institutional asset-management operating loop can act on the measurements created by Atlas and LAIS.

30.2 Institutional decision loop

A mature implementation can be expressed as:

Identify → Verify → Normalize → Mark → Structure → Size → Route → Monitor → Re-Mark → Finance/Transfer → Realize → Learn

The relationship is:

LAIS establishes trust. Atlas establishes measurement. The institutional operating loop establishes action. The asset manager supplies capital allocation, operating intervention, and investment judgment.

30.3 Institutional operating mapping

30.4 Asset-centric economics

Institutional asset management begins with the asset rather than reported company financials.

Financial statements remain important for operating-company analysis, liquidity, covenants, servicing capacity, and enterprise risk. But for a legal-asset platform, economic truth ultimately depends on the underlying assets: their authenticity, rights, expected recovery, timing, concentration, servicing, and collectability.

The operating principle is therefore:

Everything starts with a verified asset state and ends with realized asset cash flow.

This is especially important in acquisitions. An originator's historical EBITDA may not reveal weak cases, stale assets, hidden lien disputes, duration extension, or a concentration problem. Asset-level reconstruction can.

31. The Institutional Decision Genome

31.1 Institutional judgment should compound

An experienced asset manager creates value not only because it owns data, but because it repeatedly makes decisions under uncertainty and learns where judgment differs from market convention.

The Institutional Decision Genome is a proposed system for capturing that process.

For each material decision, store:

Information → Model → Rule → Judgment → Exception → Structure → Capital Source → Decision → Outcome

31.2 Minimum Decision Genome record

Each record should contain:

• —decision date;
• —asset state and evidence snapshot;
• —LAIS and ECS;
• —model version;
• —economic mark distribution;
• —duration distribution;
• —portfolio state;
• —applicable institutional rules;
• —Atlas recommendation;
• —final human decision;
• —override rationale;
• —approved structure;
• —capital source;
• —surveillance requirements;
• —realized cash flows;
• —realized duration;
• —integrity events; and
• —attribution of gain or loss.

31.3 Decision Alpha

The institution should measure whether discretionary decisions added value relative to a defined benchmark.

An illustrative metric is:

\[ DecisionAlpha_i = RealizedEconomicOutcome_i - BenchmarkExpectedOutcome_i \]

Aggregate attribution can then test whether value came from:

• —sourcing;
• —rejecting weak assets;
• —pricing;
• —structuring;
• —duration judgment;
• —capital routing;
• —portfolio construction;
• —servicing intervention;
• —refinancing;
• —exit timing; or
• —human override.

31.4 Override alpha

For assets where humans override the model:

\[ OverrideAlpha = Outcome_{override} - CounterfactualOutcome_{modelpolicy} \]

The counterfactual is imperfect and should be handled carefully, but tracking override cohorts can still reveal whether discretionary intervention is systematically beneficial.

The purpose is not to replace expert investment judgment. It is to make successful judgment reproducible across the institution.

31.5 Rule formation

An institutional rule should graduate from observation to doctrine through a controlled pathway:

1. 1.repeated pattern observed;
2. 2.retrospective evidence assembled;
3. 3.economic significance tested;
4. 4.rule proposed;
5. 5.rule challenged;
6. 6.decision authority approves;
7. 7.rule enters shadow mode;
8. 8.production impact measured;
9. 9.rule activated, revised, or rejected; and
10. 10.outcomes continuously monitored.

This prevents institutional lore from becoming automated policy without evidence.

32. Information Seasoning and the Separation of Asset Duration From Capital Duration

32.1 The central capital-management insight

A legal asset may require years to reach final resolution. Retained manager capital does not necessarily need to remain exposed for the same period.

\[ LegalAssetDuration \neq RetainedCapitalDuration \]

If uncertainty decreases as an asset seasons, the asset can become easier to finance, syndicate, insure, refinance, or transfer before legal resolution.

32.2 Information seasoning

Information seasoning is the process by which the quality of a legal asset's observable state improves over time.

Examples include:

• —verified procedural progression;
• —additional liability evidence;
• —treatment completion;
• —clearer lien position;
• —improved settlement visibility;
• —stable counsel servicing behavior;
• —reduction in contradictory records;
• —resolution of priority disputes; and
• —repeated successful servicing updates.

The information state may therefore improve even while the case remains unresolved.

32.3 Information Gain Score

A useful institutional metric is Information Gain, measuring how much uncertainty has been reduced since acquisition or funding.

Illustratively:

\[ IG_t = H(\Theta|D_0)-H(\Theta|D_t) \]

where \(H\) represents uncertainty about the economic state \(\Theta\), \(D_0\) is the initial evidence, and \(D_t\) is the current evidence.

In practice, the institution may use simpler proxies such as narrowing prediction intervals, higher ECS, reduced lien uncertainty, improved duration confidence, or fewer unresolved exceptions.

32.4 Financeability as a state variable

Atlas should maintain a Financeability State separate from raw asset quality.

Inputs can include:

• —LAIS;
• —ECS;
• —seasoning period;
• —payment or servicing history;
• —documentation completeness;
• —portfolio concentration;
• —volatility of mark;
• —duration uncertainty;
• —lender eligibility rules;
• —advance-rate rules; and
• —unresolved exceptions.

An asset can be economically attractive but not yet financeable. As information improves, capital treatment may change.

33. Capital Routing, Capital Velocity, and Institutional Economics

33.1 The objective is not to own every dollar of exposure

A sophisticated asset manager distinguishes between attractive exposure and the optimal source of capital for that exposure.

Possible routes include:

• —manager balance sheet;
• —warehouse facility;
• —acquisition financing;
• —syndicated participation;
• —third-party institutional capital;
• —insured or reinsured structure;
• —securitized or pooled financing;
• —strategic partner capital; and
• —secondary transfer.

The operating system should determine the best capital route subject to eligibility, return, liquidity, control, and risk constraints.

33.2 Capital Velocity Ratio

A proposed institutional metric is:

\[ CVR = \frac{Annualized\ Gross\ Asset\ Exposure\ Created}{Average\ Retained\ Capital} \]

CVR should never be maximized without regard to risk. Its purpose is to expose whether information, structuring, and financing allow the asset manager to create more high-quality asset exposure per dollar of retained capital.

33.3 Risk-adjusted capital productivity

A stronger metric is:

\[ RACP = \frac{Expected\ Economic\ Value\ Added - Expected\ Loss - Operating\ Cost}{Average\ Retained\ Capital} \]

This shifts the optimization target from headline volume to economic value created per dollar of scarce capital.

33.4 Capital release events

Atlas should record which events cause an asset to become eligible for improved capital treatment.

Examples:

• —LAIS crosses a defined threshold;
• —ECS improves after independent verification;
• —key lien dispute is cured;
• —duration interval narrows;
• —case reaches a defined procedural milestone;
• —collateral perfection is completed;
• —servicing history reaches a threshold;
• —concentration falls after portfolio growth; or
• —a lender rule is satisfied.

This turns diligence and servicing into measurable capital-formation activities.

33.5 Staged commitment gates

The institutional operating system should treat capital commitment as a sequence of evidence-gated decisions rather than a binary yes/no event. The governing principle is:

Commit only the amount of capital justified by the current evidence state, then earn the right to increase, refinance, or release capital as uncertainty resolves.

An illustrative gate structure is:

The gate model formalizes staged commitment, downside protection, capital recovery, and the distinction between asset ownership and capital control.

33.6 Capital routing should be explainable

Every capital-routing recommendation should identify:

• —eligible funding sources;
• —expected return to each source;
• —retained risk;
• —expected capital duration;
• —concentration impact;
• —covenant or facility constraints;
• —liquidity impact;
• —downside exposure; and
• —reason the recommended route dominates alternatives.

This prevents capital allocation from becoming another black box.

34. The Roll-Up as a Quantitative Learning Flywheel

34.1 Acquisitions should be treated as data and control events

A roll-up can create scale without creating intelligence. The institutional objective is different: each acquisition should make the system measurably smarter and more financeable.

The flywheel is:

Acquire → Reconstruct → Normalize → Verify → Re-Mark → Detect Leakage → Apply Institutional Rules → Improve Servicing → Finance → Observe Outcomes → Learn → Improve Acquisition Underwriting

34.2 Day-one book reconstruction

Every acquired book should be processed through a standardized Day-One protocol:

1. 1.source-system inventory;
2. 2.data-rights verification;
3. 3.schema mapping;
4. 4.entity resolution;
5. 5.Legal Asset Passport creation;
6. 6.case verification;
7. 7.lien and collateral reconstruction;
8. 8.LAIS and ECS scoring;
9. 9.economic re-marking;
10. 10.duration re-estimation;
11. 11.portfolio concentration mapping;
12. 12.integrity-exception creation;
13. 13.servicing triage;
14. 14.financing-eligibility mapping; and
15. 15.acquisition thesis variance analysis.

34.3 Acquisition value bridge

For each acquisition, Atlas should reconcile:

\[ PurchasePrice \rightarrow VerifiedAssetValue \rightarrow IntegrityAdjustments \rightarrow ServicingUplift \rightarrow FinancingUplift \rightarrow RealizedValue \]

This allows the asset manager to distinguish value created through buying well from value created after acquisition.

34.4 Data moat with governance

Acquisitions can compound proprietary information across:

• —cases;
• —counsel;
• —providers;
• —liens;
• —servicing behavior;
• —settlement outcomes;
• —financing outcomes;
• —capital structures; and
• —acquisition performance.

But the data moat should never be framed simply as quantity. The institutional moat is governed, permissioned, normalized, outcome-linked data with reproducible lineage.

That is harder to copy than raw records.

35. Legal Asset Research Laboratory and Institutional Benchmarks

35.1 From operating data to market infrastructure

Once sufficient breadth exists, the platform can produce anonymized institutional benchmarks that the litigation finance market currently lacks.

Potential indices include:

• —Legal Asset Duration Index;
• —Venue Recovery Index;
• —Counsel Reliability Distribution;
• —Medical Lien Integrity Index;
• —Portfolio Integrity Migration Index;
• —Settlement Timing Index;
• —Legal Asset Financeability Index;
• —Servicing Quality Index; and
• —Integrity Event Frequency Index.

35.2 Benchmark governance

Any public benchmark should have:

• —published definition;
• —inclusion/exclusion rules;
• —minimum sample requirements;
• —revision policy;
• —methodology version;
• —survivorship-bias controls;
• —confidentiality protections;
• —conflict disclosures; and
• —independent methodology review.

A benchmark becomes strategically valuable only when market participants trust how it is constructed.

35.3 State of Legal Asset Integrity

The annual State of Legal Asset Integrity report should eventually report aggregate, anonymized statistics such as:

• —distribution of LAIS scores;
• —prevalence of major integrity exceptions;
• —evidence-confidence trends;
• —duration drift;
• —lien and duplicate-funding signals;
• —concentration trends;
• —model calibration;
• —servicing quality;
• —material regulatory changes; and
• —observed relationships between integrity and realized economics.

This can establish thought leadership without disclosing proprietary underwriting alpha.

36. Institutional Proof of the Operating Model

36.1 The system should be judged by economic outcomes

The operating model succeeds only if it improves real decisions and real capital outcomes.

The core scorecard should include:

Asset quality

• —MIF frequency;
• —realized loss;
• —LAIS migration;
• —exception cure rate;
• —evidence-confidence improvement.

Decision quality

• —mark calibration;
• —duration calibration;
• —false-negative integrity cost;
• —Decision Alpha;
• —override performance;
• —pass-rate quality.

Operating quality

• —time from intake to IC-ready decision;
• —percentage of diligence automated;
• —evidence freshness;
• —manual touches per asset;
• —servicing exception resolution time.

Portfolio quality

• —concentration-adjusted return;
• —expected shortfall;
• —liquidity-at-risk;
• —portfolio integrity distribution;
• —vintage dispersion.

Capital quality

• —advance-rate improvement;
• —facility eligibility improvement;
• —capital release time;
• —Capital Velocity Ratio;
• —risk-adjusted capital productivity;
• —external capital participation rate.

36.2 The ultimate operating-model experiment

The strongest demonstration is a controlled historical and prospective comparison:

1. 1.reconstruct historical institutional decisions;
2. 2.run Atlas in frozen-date replay;
3. 3.identify where each outperformed;
4. 4.encode proven institutional rules;
5. 5.operate Atlas in shadow mode on live opportunities;
6. 6.compare human-only, model-only, and combined decisions;
7. 7.measure realized economics; and
8. 8.promote only statistically and economically supported rules into production.

The target is not artificial intelligence for its own sake.

The target is a recursive institutional learning system that compounds asset knowledge, decision quality, and capital efficiency over time.

37. Institutional Architecture Summary

The full architecture can be summarized as follows:

RAW LEGAL ECOSYSTEM DATA
        |
        v
DATA RIGHTS + PROVENANCE + ENTITY RESOLUTION
        |
        v
LEGAL ASSET PASSPORT / STATE VECTOR
        |
        +------------------------------+
        |                              |
        v                              v
LAIS INTEGRITY SCORE            EVIDENCE CONFIDENCE
        |                              |
        +---------------+--------------+
                        |
                        v
             PROBABILISTIC ECONOMIC MARK
                 + DURATION DISTRIBUTION
                        |
                        v
               the asset manager DECISION GENOME
                        |
                        v
                 PORTFOLIO DIGITAL TWIN
                        |
                        v
              STRUCTURE + CAPITAL ROUTING
                        |
                        v
              SERVICING + SURVEILLANCE
                        |
                        v
              RE-MARK + FINANCE/TRANSFER
                        |
                        v
                 REALIZED OUTCOMES
                        |
                        v
             MODEL + RULE LEARNING LOOP
                        |
                        +---------------------> back to decisioning

The strategic separation is deliberate:

• —LAIS defines how legal assets should be trusted.
• —Atlas measures their state, uncertainty, value, and evolution.
• —The institutional operating loop determines what action should follow.
• —The asset manager contributes capital allocation, operating intervention, structuring, financing, and experienced investment judgment.

This architecture converts legal finance from a collection of individually underwritten transactions into a quantitatively managed asset system.

Part IV. Capital Markets, Implementation, and Industry Leadership

38. Institutional Capital Markets Implications

Legal Asset Integrity can change more than underwriting. It can improve how legal assets are financed.

38.1 Warehouse lending

Lenders need confidence that collateral exists, is properly pledged, is not duplicated, and is being serviced according to agreed standards.

LAIS can support:

• —borrowing-base eligibility;
• —concentration limits;
• —collateral audit prioritization;
• —delinquency / integrity triggers;
• —dynamic advance rates;
• —covenant monitoring.

38.2 Asset-backed financing and securitization

Standardized asset records and lifecycle data can support:

• —consistent collateral tapes;
• —stratification tables;
• —exception reporting;
• —vintage analysis;
• —servicing performance metrics;
• —loss and duration curves;
• —independent review;
• —investor transparency.

The standard does not itself make a legal asset securitizable. It can, however, reduce the information asymmetry that makes structured financing difficult.

38.3 Insurance and risk transfer

Verified data can improve discussions around:

• —adverse judgment risk;
• —portfolio protection;
• —after-the-event or judgment-preservation products where applicable;
• —credit enhancement;
• —reinsurance or structured risk transfer.

38.4 Acquisition financing

A buyer can use LAIS to quantify the integrity of an acquisition target's portfolio before assigning value.

This creates a direct bridge between operational diligence and purchase price.

38.5 Public-market readiness

A public or public-market-adjacent platform must explain its assets, controls, data, conflicts, model governance, and portfolio performance to investors who did not originate the assets.

A standardized integrity framework creates a more credible narrative:

"We do not simply acquire legal assets. We verify them, score their integrity, monitor them continuously, govern model-driven decisions, and maintain auditable evidence across the lifecycle."

That is a materially different proposition from a collection of funding companies operating on separate spreadsheets and local judgment.

39. Implementation Roadmap

Phase 1: Establish the control spine (0 to 90 days)

Objectives

• —define canonical asset schema;
• —deploy Legal Asset Passport;
• —implement jurisdiction eligibility gate;
• —establish CaseVerify minimum standard;
• —create hard-stop taxonomy;
• —create LAIS v1 and Evidence Confidence Score;
• —inventory existing portfolios and data sources;
• —establish model registry and exception log.

Success measures

• —100% of new assets receive unique global asset IDs;
• —100% of new transactions pass jurisdiction gate;
• —minimum evidence completeness threshold established;
• —hard stops are machine-detectable where data permits;
• —every production model is registered.

Phase 2: Normalize and surveil (90 to 180 days)

Objectives

• —ingest existing portfolio books;
• —entity-resolve law firms, attorneys, providers, defendants, and insurers;
• —launch Attorney Risk Graph;
• —integrate lien / filing checks;
• —launch Portfolio Surveillance;
• —implement data freshness rules;
• —begin continuous LAIS rescoring.

Success measures

• —majority of portfolio assets converted to Passport format;
• —concentration views available across core counterparties;
• —stale-case alerts operational;
• —exception aging and cure workflow measured;
• —score migration available by vintage.

Phase 3: Protect settlement and scale governance (6 to 12 months)

Objectives

• —deploy SettlementLock;
• —automate payoff and waterfall verification;
• —implement Disclosure and Governance Module;
• —integrate model drift monitoring;
• —establish integrity committee;
• —publish internal quarterly Legal Asset Integrity report.

Success measures

• —material settlement events reconciled to payoff workflow;
• —payment instruction changes independently verified;
• —jurisdiction-rule changes mapped to affected assets;
• —model overrides measured and reviewed;
• —lender/investor integrity reporting standardized.

Phase 4: Institutionalize the operating loop and capital routing (12 to 18 months)

Objectives

• —deploy frozen-date replay across historical institutional decisions;
• —operate model-only, human-only, and combined decision comparisons;
• —establish Institutional Decision Genome records;
• —quantify Decision Alpha and override performance;
• —launch portfolio digital twin and stress-testing program;
• —implement staged commitment gates;
• —map assets to warehouse, syndication, insurance, and external-capital eligibility;
• —measure Information Gain and capital-release events;
• —create acquisition Day-One book reconstruction as a repeatable operating protocol.

Success measures

• —historical decision replay completed for a representative set of historical assets and transactions;
• —model and override performance reported by cohort;
• —material new investments carry an explicit capital-routing rationale;
• —capital-duration distribution is measured separately from asset-duration distribution;
• —financing eligibility and advance-rate changes are linked to observable information improvements;
• —acquisition integration produces a verified book-level state within a defined service level.

Phase 5: Externalize the standard (18 to 24 months)

Objectives

• —publish LAIS methodology principles;
• —form Legal Asset Integrity Council;
• —create external certification / audit framework;
• —launch permissioned Legal Asset Registry network;
• —publish annual State of Legal Asset Integrity report;
• —create benchmark data products.

Success measures

• —external participants adopt common integrity fields;
• —lender or investor counterparties reference LAIS in diligence;
• —registry detects cross-platform duplicate interests;
• —industry benchmark becomes a recognized reference point;
• —at least one external financing or diligence process consumes standardized Legal Asset Passport fields.

40. Measuring Whether the Standard Works

A thought-leadership standard must be falsifiable. It should define success metrics.

40.1 Integrity KPIs

• —percentage of assets independently case-verified;
• —percentage with clear lien/payoff mapping;
• —percentage with current counsel verification;
• —percentage with ECS above threshold;
• —hard-stop incidence;
• —exception cure time;
• —stale-data rate;
• —duplicate-funding detection rate;
• —unresolved priority-conflict rate;
• —settlement reconciliation time;
• —payment instruction exception rate.

40.2 Portfolio KPIs

• —LAIS distribution;
• —LAIS migration;
• —ECS distribution;
• —integrity risk by originator;
• —integrity risk by attorney;
• —integrity risk by provider;
• —concentration breaches;
• —duration drift;
• —reserve changes attributable to new evidence;
• —realized loss by integrity band.

40.3 Model-governance KPIs

• —calibration by model and cohort;
• —false positive / false negative rate for anomaly models;
• —model drift incidents;
• —human override rate;
• —override performance;
• —percentage of scores reproducible from stored evidence;
• —production models without current validation.

40.4 Institutional outcome KPIs

Over time, the strongest evidence of value should include:

• —lower fraud and operational loss;
• —fewer collateral disputes;
• —faster acquisition diligence;
• —lower servicing cost per asset;
• —better lender reporting;
• —improved borrowing terms where supported by performance;
• —lower exception frequency;
• —stronger recovery predictability;
• —faster identification of deteriorating assets.

41. Illustrative Use Cases

The following scenarios are intentionally simplified. They illustrate how the standard operates, not how any specific investment should be decided.

Use Case A: Duplicate Funding

A claimant is submitted by a law firm for a new advance. The case is authentic and economically attractive.

CaseVerify confirms the docket and counsel. The Legal Asset Registry identifies an existing funded interest under a name variant. The prior funder is not disclosed in the intake file.

Result: Collateral Clarity triggers a hard stop. The transaction is not approved until the existing obligation, priority, and available proceeds are reconciled.

Lesson: A strong case is not a clean asset if the economic rights are unclear.

Use Case B: Stale Case Presented as Active

An acquired portfolio reports a matter as "active litigation." Docket monitoring shows no activity for fourteen months and identifies a dismissal event that was not reflected in servicing records.

Result: Case Authenticity and Servicing Risk downgrade immediately. The asset is routed for manual review and reserve reassessment.

Lesson: Portfolio surveillance must reconcile servicing narratives against external evidence.

Use Case C: Medical Provider Outlier

A provider appears across a rapidly growing share of cases from several attorneys. Treatment intensity and lien amounts materially exceed comparable cohorts, while realized settlement values do not increase proportionally.

Result: Medical Lien Integrity and the relationship graph flag the provider cluster. New exposure is subjected to enhanced diligence and concentration limits.

Lesson: A graph-level pattern can reveal risk invisible at the individual case level.

Use Case D: Settlement Payment Instruction Change

A case is reported settled. Hours before payment, an email requests that payoff funds be wired to a different account.

Result: SettlementLock blocks the change pending independent verification through a known contact and verified beneficiary process.

Lesson: Strong underwriting does not eliminate payment fraud risk.

Use Case E: Attractive Economics, Weak Evidence

A high-value case receives strong projected recovery and return metrics. LAIS is 81, but Evidence Confidence is only 48 because core information is self-reported and several external sources are stale.

Result: The platform does not approve based on economics. It requires evidence cure before capital deployment.

Lesson: Uncertainty should be explicit and decision-relevant.

Use Case F: Acquisition Book Triage

A buyer receives a 20,000-asset portfolio from an acquisition target. Manual re-underwriting would take months.

Atlas ingests the tape, resolves entities, verifies dockets, reconstructs evidence freshness, scores LAIS/ECS, identifies duplicate interests, and segments the portfolio into:

• —verified / standard review;
• —conditionally verified;
• —high-priority remediation;
• —hard-stop / impaired.

Result: Human diligence focuses first on assets most likely to affect valuation.

Lesson: Integrity infrastructure turns acquisition diligence into a scalable portfolio process.

42. An Industry Standard Without Giving Away the Moat

For a standard to become influential, the market must be able to understand and adopt it. For a technology platform to retain competitive advantage, it should not publish every model, feature, threshold, and data relationship.

The recommended strategy is open standard, proprietary intelligence.

Public / industry standard

Publish:

• —LAIS definitions;
• —eight pillars;
• —minimum data categories;
• —evidence hierarchy;
• —hard-stop principles;
• —control and governance requirements;
• —confidence methodology principles;
• —audit expectations;
• —certification framework.

Proprietary layer

Protect:

• —feature engineering;
• —model weights by asset class;
• —anomaly detection logic;
• —entity graph;
• —proprietary outcome data;
• —calibration methods;
• —partner data combinations;
• —fraud patterns;
• —real-time rules;
• —portfolio optimization algorithms.

This allows the standard to become industry language while the implementation remains differentiated.

43. Building the Thought-Leadership Platform

The standard should not be a one-time white paper. It should become an institution around the category.

43.1 The Legal Asset Integrity Council

Create an independent advisory council including perspectives from:

• —litigation finance;
• —plaintiff and defense bars;
• —former judges;
• —legal ethics;
• —structured finance;
• —servicing;
• —risk and compliance;
• —data governance;
• —insurance;
• —institutional investors.

The council should challenge the standard, not simply endorse it.

43.2 Annual State of Legal Asset Integrity Report

Publish anonymized market-level findings such as:

• —verification failure rates;
• —stale case rates;
• —lien conflicts;
• —duplicate funding signals;
• —duration drift;
• —settlement reporting lag;
• —counsel concentration;
• —provider concentration;
• —exception frequency;
• —integrity score migration;
• —regulatory developments.

This can create a benchmark the market lacks today.

43.3 LAIS certification

Over time, establish a certification process for:

• —funding platforms;
• —servicers;
• —portfolio tapes;
• —warehouse collateral pools;
• —acquired books;
• —data and model governance.

Certification should require evidence and testing, not payment for a badge.

43.4 Research program

Priority research topics should include:

• —relationship between integrity score and realized loss;
• —duration prediction by venue and case type;
• —duplicate funding prevalence;
• —servicing behavior and recovery outcomes;
• —provider concentration and settlement economics;
• —attorney update reliability;
• —impact of evidence freshness on underwriting error;
• —model explainability and decision quality.

43.5 Policy engagement

The strongest policy posture is neither "all funding is good" nor "all regulation is bad."

It is:

Access to capital and integrity are complements. A transparent, well-governed funding market can preserve access while reducing the behaviors that drive regulatory backlash.

That is a more durable thought-leadership position.

44. Strategic Implications for the Litigation Finance Industry

The litigation finance market is likely to consolidate around platforms that can combine capital, origination, servicing, data, technology, and governance.

The winner will not simply be the firm with the largest balance sheet.

The durable advantage will come from a recursive system:

1. 1.acquire or originate assets;
2. 2.verify and normalize them;
3. 3.underwrite them with transparent models;
4. 4.monitor them continuously;
5. 5.observe realized outcomes;
6. 6.learn which rules and signals were predictive;
7. 7.improve future acquisition, pricing, and servicing;
8. 8.use stronger evidence to attract lower-cost and more scalable capital.

This creates a compounding data and operating moat.

The strategic shift is from:

funding lawsuits

To:

managing verified legal assets as an institutional portfolio.

That is the category transition the Legal Asset Integrity Standard is designed to accelerate.

45. Limitations and Responsible Use

A credible standard must state what it cannot do.

45.1 No scoring system eliminates fraud

Sophisticated fraud adapts. Integrity technology reduces risk by increasing verification, detecting anomalies, improving accountability, and shortening detection time. It cannot guarantee that deception will never occur.

45.2 Data can be incomplete or biased

Court coverage varies. Settlements may be confidential. Outcomes can be selectively reported. Medical data may be incomplete. Historical data may reflect prior underwriting bias.

Sparse-data warnings and evidence confidence are therefore mandatory.

45.3 Correlation is not misconduct

A relationship graph can identify unusual patterns. It cannot by itself establish fraud, unethical behavior, or legal liability. Adverse actions should require appropriate review and evidence.

45.4 Models do not replace counsel

Atlas and LAIS support financial and integrity decisions. They do not provide legal representation or replace attorney professional judgment.

45.5 Jurisdiction rules change

The regulatory rules engine requires continuous legal maintenance. A rule set is only as reliable as its latest validated update.

45.6 Standards require governance

Publishing LAIS without enforcing exceptions, data quality, model discipline, and independent review would create the appearance of control without the substance of control.

46. Conclusion

Litigation finance is at an inflection point.

The market has demonstrated that legal claims, law-firm receivables, medical liens, and other litigation-linked interests can attract sophisticated capital. The next question is whether those assets can be managed with the verification, surveillance, transparency, and governance expected of an institutional asset class.

The answer cannot be another layer of manual review.

It requires infrastructure.

The Legal Asset Integrity Standard establishes a framework in which:

• —every asset has a verifiable identity;
• —every economic interest is mapped;
• —every material fact has source lineage;
• —uncertainty is explicit;
• —every critical exception has an owner;
• —every asset is monitored over time;
• —capital rights remain separate from litigation control;
• —every model can be explained and reproduced;
• —every settlement can be reconciled; and
• —every realized outcome improves the next decision.

LAIS provides the common scoring language.

The Legal Asset Passport provides the canonical record.

The Evidence Confidence Score makes uncertainty visible.

The Integrity Event Ledger provides continuous history.

The eight Atlas product modules provide an operating architecture capable of implementing the standard at scale.

The broader objective is not to make litigation finance appear safer. It is to make it measurably safer, more transparent, more governable, and more financeable.

The industry does not need to choose between access to capital and institutional discipline.

It can build both.

The future of litigation finance will not be defined by who deploys the most capital. It will be defined by who can prove the integrity of the legal assets behind that capital.

Appendix A: LAIS Summary Scorecard

Appendix B: Illustrative Integrity Event Codes

Appendix C: Illustrative Hard-Stop Codes

Appendix D: Sample Legal Asset Integrity Report

Asset

• —Asset ID: [redacted]
• —Jurisdiction: [state / court]
• —Case type: [type]
• —Last verified: [timestamp]

Decision gates

• —Jurisdiction eligibility: Eligible / Conditional / Prohibited
• —LAIS: 84
• —Evidence Confidence: 92
• —Hard stops: None
• —Active warnings: 2

Pillar scores

Top integrity drivers

1. 1.Court and counsel identity independently verified.
2. 2.Lien acknowledgment present and payoff chain documented.
3. 3.No duplicate funding match found within available registry sources.
4. 4.Provider concentration above cohort median.
5. 5.Case duration modestly above comparable venue band.

Required actions

• —Refresh provider lien documentation within 30 days.
• —Increase docket monitoring frequency until next material event.

Governance

• —Model versions recorded.
• —No human override.
• —Evidence lineage complete.
• —Next automatic review: [date/event].

Appendix E: Recommended Portfolio Integrity Dashboard

A portfolio-level report should prioritize action rather than visual complexity.

Page 1: Executive exceptions

• —assets with new hard stops;
• —largest LAIS downgrades;
• —unresolved payment exceptions;
• —jurisdiction-rule impacts;
• —assets with low Evidence Confidence;
• —concentration limit breaches;
• —settlements requiring reconciliation.

Page 2: Portfolio integrity distribution

• —LAIS by band;
• —ECS by band;
• —score migration 30 / 90 / 180 days;
• —exception aging;
• —stale evidence rate.

Page 3: Concentration

• —counsel;
• —provider;
• —venue;
• —originator;
• —defendant / insurer;
• —case type;
• —duration bucket.

Page 4: Learning

• —realized loss by LAIS band;
• —realized duration vs expected;
• —realized recovery vs expected;
• —override performance;
• —fraud / integrity incidents by root cause;
• —rules changed from observed outcomes.

Appendix F: Publication and Stewardship Model

The standard should be maintained through versioned releases.

Recommended governance

• —Version 1: founding methodology
• —Annual public review
• —Interim updates for material regulatory or technical developments
• —External advisory council review
• —Public change log
• —Separate proprietary Atlas implementation releases

Recommended public artifacts

1. 1.Legal Asset Integrity Standard white paper
2. 2.LAIS methodology overview
3. 3.Annual State of Legal Asset Integrity report
4. 4.Quarterly regulatory integrity update
5. 5.Technical note on explainable legal asset models
6. 6.Integrity certification framework
7. 7.Public glossary / data dictionary

Appendix G: Minimum Institutional Model Validation Report

Every Tier 1 model should have a validation report with, at minimum:

1. 1.model identifier and version;
2. 2.business purpose and decisions influenced;
3. 3.model owner and independent validator;
4. 4.development sample and observation window;
5. 5.target definition and estimand;
6. 6.feature families and data-rights status;
7. 7.leakage controls and decision-time boundary;
8. 8.missing-data treatment;
9. 9.benchmark and challenger specifications;
10. 10.discrimination metrics;
11. 11.calibration metrics;
12. 12.error-cost analysis;
13. 13.subgroup and transfer analysis;
14. 14.stability and drift testing;
15. 15.sensitivity analysis;
16. 16.known limitations;
17. 17.prohibited uses;
18. 18.override implications;
19. 19.monitoring thresholds;
20. 20.validation conclusion and conditions of use; and
21. 21.approval record.

A validation conclusion should be one of:

• —approved;
• —approved with conditions;
• —restricted use;
• —remediation required; or
• —rejected.

Appendix H: Institutional Decision Record

A material decision should be reproducible from one structured record.

Decision context

• —opportunity ID;
• —decision date/time;
• —decision type;
• —acquisition/funding/financing/servicing/exit classification;
• —responsible committee or authority.

Asset state

• —Legal Asset Passport version;
• —LAIS;
• —ECS;
• —hard stops;
• —warnings;
• —unresolved exceptions;
• —jurisdiction eligibility.

Economics

• —P10/P50/P90 recovery;
• —expected mark;
• —duration distribution;
• —expected loss;
• —IRR/MOIC distribution;
• —downside stress results.

Portfolio effect

• —marginal concentration;
• —liquidity-at-risk;
• —expected-shortfall impact;
• —correlated-exposure flags;
• —facility impact.

Doctrine and judgment

• —applicable institutional rules;
• —Atlas recommendation;
• —human decision;
• —override reason;
• —evidence cited;
• —required covenants or conditions.

Capital routing

• —initial capital source;
• —retained capital;
• —financing plan;
• —capital-release gates;
• —expected retained capital duration.

Outcome

• —realized recovery;
• —realized duration;
• —realized loss;
• —integrity events;
• —servicing interventions;
• —financing events;
• —Decision Alpha attribution;
• —lessons promoted to candidate institutional rules.

Appendix I: Day-One Acquisition Book Reconstruction Protocol

For every portfolio or operating-company acquisition, the following minimum controls should be completed or explicitly waived through governance:

Data and rights

• —inventory source systems;
• —verify data-use rights;
• —preserve original snapshots;
• —generate ingestion hashes;
• —map source fields to canonical schema.

Asset identity

• —assign global asset IDs;
• —resolve claimant/case/counsel entities;
• —verify docket and status;
• —identify duplicates;
• —identify missing or orphaned records.

Economic rights

• —reconstruct contract rights;
• —map liens and priority;
• —reconcile principal and accruals;
• —verify payoff rights;
• —classify unresolved title/priority issues.

Quantitative state

• —calculate LAIS;
• —calculate ECS;
• —produce mark distribution;
• —produce duration distribution;
• —identify integrity exceptions;
• —map concentration.

Servicing

• —reconcile last update;
• —test contact and documentation SLAs;
• —identify stale assets;
• —identify settlements not fully reconciled;
• —prioritize highest expected-loss exceptions.

Capital treatment

• —map facility eligibility;
• —identify financeable subsets;
• —identify assets requiring cure;
• —estimate retained capital need;
• —establish capital-release roadmap.

Acquisition thesis reconciliation

• —compare seller tape with verified state;
• —quantify integrity haircut;
• —quantify mark variance;
• —quantify duration variance;
• —identify value creation from servicing and financing;
• —establish post-close baseline for realized attribution.

Appendix J: Quantitative Definitions

References

1. 1.U.S. Government Accountability Office, Third-Party Litigation Financing: Market Characteristics, Data, and Trends, GAO-23-105210, December 20, 2022.
2. 2.U.S. Government Accountability Office, Intellectual Property: Information on Third-Party Funding of Patent Litigation, GAO-25-107214, December 2024.
3. 3.North Carolina General Assembly, Session Law 2026-14, House Bill 315, Prohibit Litigation Investments Act, approved June 22, 2026.
4. 4.U.S. Senate Committee on the Judiciary, Grassley Proposes Third-Party Litigation Funding Reform, Foreign Reporting Requirements, February 11, 2026.
5. 5.American Bar Association, What Attorneys Should Consider When Thinking About Third-Party Litigation Funding, Law Practice Today, December 2025, and related ABA professional responsibility materials addressing attorney independence, confidentiality, privilege, settlement authority, and disclosure.
6. 6.Board of Governors of the Federal Reserve System, Office of the Comptroller of the Currency, and Federal Deposit Insurance Corporation, Revised Guidance on Model Risk Management, April 17, 2026.
7. 7.Basel Committee on Banking Supervision, Principles for Effective Risk Data Aggregation and Risk Reporting (BCBS 239 / SRP36), current Basel Framework.
8. 8.National Institute of Standards and Technology, Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1, January 26, 2023, together with subsequent NIST AI RMF resources and profiles.
9. 9.International Organization for Standardization, ISO/IEC 42001:2023, Information Technology, Artificial Intelligence, Management System.
10. 10.U.S. Securities and Exchange Commission, Asset-Backed Securities Disclosure and Registration, Regulation AB II, Release No. 33-9638, September 4, 2014, and current SEC staff interpretations regarding asset-level disclosure.
11. 11.Bank for International Settlements, Risk Data Aggregation and Risk Reporting, including the Basel Committee's current framework and 2026 implementation materials.

Questions on the Standard

Institutional buyers, capital partners, and governance teams evaluating the Legal Asset Integrity Standard.