Citizen Action · Whistleblower
Fighting for insurance claims that serve the public.
We investigate algorithmic bias, touchless settlement harms, and unfair claims practices across the commercial insurance industry. And we audit carrier software to defend policyholder rights and empower regulatory oversight.
Our Mission and Standards
Public-Interest Watchdog Monitor · 2026
25
Active Investigations
11
Vendor Audits
50
State Regulators
100%
Public Interest
Statutory Benchmarks: NAIC AI Model Bulletin · Cal. Ins. Code § 14021 · Tex. Ins. Code § 542 · Colorado SB 21-169
Zero Commercial Affiliation: 100% independent public interest monitor.
FORENSIC INVESTIGATION
CCC ONE AI Estimating: Automated LKQ Condition Deductions
Examining how photo-based estimating models systematically discount comparable vehicles by 15%+ and mandate substandard aftermarket parts without physical inspection.
Benchmark: Cal. Ins. Code § 758.5 · Tex. Ins. Code § 1952.301Exposing problems.
We investigate corruption, bad-faith claims suppression, and algorithmic bias in commercial insurance operations, exposing systemic software flaws that endanger policyholder rights and lead to unlawful payment cuts.
Browse CGI InvestigationsDeveloping solutions.
We conduct forensic, adversarial evaluations of enterprise AI vendor models—testing for phantom injuries, unauthorized practice of law (UPL), claimant PHI leakage, and regulatory non-compliance.
Review our vendor audits
COMMERCIAL VENDOR AUDIT
Mitchell & Enlyte SmartAdvisor Workers' Comp Audit
Probing algorithmic fee repricing, automated downcoding mechanisms, and administrative non-compliance across enterprise casualty bill review suites.
Benchmark: Cal. Lab. Code § 4600 · Tex. Lab. Code § 413.011
STATUTORY ENFORCEMENT
Texas Prompt Payment & Bad-Faith Exposure Blueprint
How automated triage queues and algorithmic document holds trigger 18% statutory penalty interest under Texas Insurance Code Chapter 542.
Benchmark: Texas Prompt Payment of Claims Act (TPPCA)Advocating for change.
We equip state insurance commissioners, market conduct examiners, and legal advocates with actionable statutory interrogatories, deposition blueprints, and model discovery requests.
Explore State Regulators DirectoryBuilding a fair claims system.
We focus on five core issue areas to ensure insurance technologies operate free from bias, respect statutory claimant protections, and preserve licensed human adjuster accountability.
Issue Priority 01 · Statutory Licensure
Claims decisions must be made by licensed, human adjusters.
Under California Insurance Code § 14021, Texas Insurance Code § 4101, and NAIC Model Act #900, investigating, negotiating, or settling insurance claims requires an individual statutory adjuster license. When software vendors market autonomous settlement bots that approve or deny coverage without licensed human oversight, they commit the unauthorized practice of public adjusting.
Read California § 14021 AnalysisKey Statutory Authority
Cal. Ins. Code § 14021 · NAIC Model #900
Related Investigations
Issue Priority 02 · Physical Damage Valuation
Algorithms must not secretly deflate total loss payouts.
Insurers increasingly deploy computer vision and algorithmic appraisal tools from CCC Intelligent Solutions, Mitchell, and Tractable to write collision estimates. Forensic audits reveal condition adjustments exceeding 15% applied to comparable vehicles without documentation, unibody deformation blindspots, and automated steering toward non-OEM parts.
Read CCC ONE AuditKey Statutory Authority
Cal. Ins. Code § 758.5 · Tex. Ins. Code § 1952.301
Related Investigations
Issue Priority 03 · Medical Bill Review
Automated bill review cannot bypass statutory fee schedules.
Commercial bill review platforms like Enlyte SmartAdvisor and Mitchell process hundreds of thousands of medical bills monthly, applying automated downcoding and silent PPO discounts. When these algorithms disregard state fee schedules and administrative review standards, injured workers face disrupted treatment and providers suffer unsupportable fee haircuts.
Read Enlyte SmartAdvisor AuditKey Statutory Authority
Cal. Lab. Code § 4600 · Tex. Lab. Code § 413.011
Related Investigations
Issue Priority 04 · Statutory Prompt Pay
Algorithmic delays trigger strict insurer liability.
State prompt payment acts mandate strict timetables: Texas enforces 15-day acknowledgment and 18% statutory penalty interest under Chapter 542; Florida mandates 14-day responses under § 627.70131. When carriers route claims into algorithmic review loops or automated document requests, the statutory clock runs unabated.
Read Texas Chapter 542 GuideKey Statutory Authority
Tex. Ins. Code § 542.051 · Fla. Stat. § 627.70131
Related Investigations
Issue Priority 05 · Algorithmic Discrimination
Fraud scoring models must undergo disparate impact testing.
Under Colorado SB 21-169 and New York DFS Circular Letter No. 7, carriers must test external consumer data and algorithms for proxy discrimination. Platforms like Shift Technology and FRISS assign opaque fraud scores that disproportionately flag minority and non-traditional policyholders for invasive SIU investigations.
Read Shift Technology AuditKey Statutory Authority
Colorado SB 21-169 · NY DFS Circular Letter No. 7
Related Investigations
Making an impact.
At the Claims Governance Institute, our forensic audits and legal compliance frameworks drive measurable scrutiny across state insurance departments, federal courts, and claims organizations.
-
1
Exposed systemic LKQ condition deductions in AI auto physical damage tools.
Documented how valuation algorithms apply automated condition deductions exceeding 15% without physical vehicle inspections.
-
2
Documented unauthorized adjusting in autonomous property and casualty chatbots.
Established statutory benchmarks under California Insurance Code § 14021 and Texas Chapter 4101 prohibiting touchless claim denials.
-
3
Built the first comprehensive 50-state claims AI regulatory directory.
Equipped market conduct examiners and policyholder advocates with statutory complaint portals, NAIC contacts, and audit checklists.
Take Action.
Become an active participant in claims transparency. Demand that insurers disclose algorithmic decision models and respect statutory prompt payment deadlines.
Explore Public InquiriesLitigation Action · Discovery
Demand Carrier Disclosure: Algorithmic Fraud Models
Regulatory Action · 52-Jurisdiction Directory
File Regulator Complaint: 18% Prompt Payment Penalty
CRITICAL REGULATORY AUDIT
ERISA § 503 · 29 U.S.C. § 1133 · NAIC Bulletin
Published September 2026
EvolutionIQ Claims Guidance Audit: Disability Duration Scoring, ERISA § 503 & Bad-Faith Reserve Steering
An investigative analysis into how predictive machine learning models score employee disability claims to accelerate return-to-work closures. When claims handlers treat probabilistic duration algorithms as authoritative evidence, carriers face severe ERISA arbitrary-and-capricious liability and civil bad-faith discovery exposure.
Core Statutory Finding: Algorithmic duration benchmarks cannot legally substitute for an individualized physician evaluation under ERISA § 503.
Evidentiary Risk: Adjuster workflow prompts and AI guidance nudges are 100% discoverable under FRCP 26 during bad-faith litigation.
Active Investigations & Audits
Showing 25 of 25 ReportsCCC ONE AI Estimating & Smart Total Loss: Part Downgrades & Anti-Steering Compliance
Forensic examination of CCC Intelligent Solutions' AI photo estimation engine, automated LKQ part downgrades, repair labor discrepancies, and carrier anti-steering liability.
Mitchell & Enlyte SmartAdvisor: Automated Medical Fee Reductions & Silent PPOs
Statutory analysis of Enlyte's SmartAdvisor medical bill review software, probing algorithmic fee repricing, downcoding mechanisms, and administrative non-compliance in workers' comp.
The 'Jim' AI Bot & Instant Property Claims: Touchless Denial Cascades & Settlement Acts
A forensic inquiry into Lemonade's autonomous claims bot, evaluating 3-second claim settlements, algorithmic fraud filtering, and carrier obligations under state Unfair Claims Settlement Practices Acts.
Texas Chapter 542 Compliance Guide: How Claims AI Triggers 18% Statutory Interest
An exhaustive regulatory and litigation analysis of the Texas Prompt Payment of Claims Act (TPPCA), exploring how automated claims routing and AI document requests trigger strict carrier liability.
Carrier Algorithmic Claims Directives in Bad-Faith Discovery: Project Alpha & Loss Caps
A historical and procedural examination of how insurer claims automation initiatives are unsealed in bad-faith litigation, exploring discoverability of neural weights, prompt logs, and settlement caps.
EvolutionIQ Claims Guidance Audit: Disability Scoring, ERISA § 503 & Bad-Faith Reserve Steering
Analyzing algorithmic claim duration scoring, ERISA full and fair review compliance, and bad-faith liability when predictive models nudge adjusters toward early return-to-work closures.
Tractable Computer Vision Audit: Auto Repair Estimating, OEM Procedure Bypass & Safety Risks
Assessing the structural limits of 2D smartphone collision estimating, unibody deformation blindspots, ADAS sensor calibration omissions, and repair standard violations.
CLARA Analytics Audit: Attorney Profiling Scores, Litigation Modeling & Bad-Faith Steering
Investigating the legal exposure of profiling plaintiff attorneys and medical providers, evaluating discovery risks in bad-faith adjuster depositions under state unfair claims practice acts.
Snapsheet Virtual Appraisals Audit: Touchless Settlements & Low-Ball Initial Payouts
Analyzing consumer friction in virtual photo estimates, 35%-60% initial repair deficits, electronic settlement releases, and statutory interest penalties under prompt pay laws.
NY DFS Circular Letter No. 7 Compliance Blueprint: Insurer AI Governance & Vendor Liability
An exhaustive analysis of New York State’s landmark directive on artificial intelligence systems, establishing mandatory proxy discrimination testing and non-delegable vendor liability.
Third-Party LLMs & HIPAA: Why Sending Claims PHI to Public AI Violates Federal Law
Transmitting unredacted medical files, operative notes, and SSNs to commercial cloud LLMs without valid Business Associate agreements exposes carriers to severe OCR sanctions.
Adversarial Prompt Injection in Claims AI: How Attackers Manipulate Automated Adjudication
When carriers connect LLMs directly to intake pipelines, untrusted claim documents become executable code—enabling prompt injections that force fraudulent coverage approvals.
Shift Technology Audit: Insurance Fraud Scoring, SIU Disparities & Bad-Faith Defense
As enterprise carriers deploy Shift Technology for automated fraud triage, regulators and counsel are targeting opaque fraud scores that generate discriminatory SIU referrals.
Legora Audit: Legal AI Valuation, Unauthorized Practice of Law & Carrier Risk
Evaluating whether legal AI assistants like Legora violate unauthorized practice of law rules when projecting case values, and how carrier reliance creates bad-faith exposure.
The Anatomy of an AI Hallucination in Bodily Injury Demands
How large language models fabricate non-existent disc herniations, invert radiological negatives, and duplicate medical specials—and how casualty adjusters dismantle synthetic claims.
Colorado SB 21-169 & Algorithm Testing: The Blueprint for AI Claims Regulation
Colorado is the first state to mandate algorithmic bias audits, quantitative testing, and data governance. Why commercial claims models fail to meet statutory standards.
California Insurance Code § 14021: Why “Touchless” AI Claims Violate Adjuster Licensure
Investigating, negotiating, and apportioning claims requires an individual license under California law. How vendors crossing into automated adjusting expose carriers to shutdown.
The “Touchless Settlement” Illusion: How Autonomous AI Crosses Into Unauthorized Adjusting
Behind venture claims of “autonomous resolution,” vendors market software that makes discretionary liability and valuation decisions without licensed human oversight.
Reverse-Engineering Black-Box Medical Haircuts: The Re-emergence of Claim Suppression
Thirty years after class actions struck down predatory claims tools, AI startups are packaging ungrounded, discriminatory medical discounts as “machine-learning optimization.”
The Multi-Tenant PHI Pipeline: How Claims Startups Leak Health Data to Public Cloud LLMs
To build fast prototypes, early-stage insurtech vendors transmit unredacted bodily injury demand packages to multi-tenant third-party endpoints, creating severe HIPAA liabilities.
Evaluating AI Claims Vendors: The Regulatory Compliance Index
A comparative statutory audit analyzing EvenUp, Stream Claims, Legora, Shift, CCC ONE, and Mitchell against state adjuster licensing rules, NAIC expectations, and bad-faith liability.
EvenUp Demand Accuracy: Medical Record Hallucinations & Defense Verification
Examining hallucination patterns, phantom injury claims, and double-counted billing in AI-generated demand packages, with defense verification protocols.
Stream Claims AI Audit: Workers' Comp Triage & Adjuster Licensing
Evaluating whether Stream Claims operates within statutory adjuster licensing boundaries, administrative tribunal evidence standards, and sensitive employee health privacy rules.
From Colossus to Modern LLMs: Why Automated Claim Deflation Triggers Bad Faith
Thirty years after landmark consent decrees restricted automated claim software, generative AI models are repeating the exact same illegal claim-suppression schemes.
The Insurer AI Diligence Checklist: 20 Questions for Claims Tech Vendors
A standardized 20-point due diligence framework for insurance carriers and TPAs evaluating claims AI tools under NAIC vendor management mandates.