AI governance and compliance failures — audit, regulatory, and traceability gaps in agent systems
46 patterns
AI governance and compliance failures — audit, regulatory, and traceability gaps in agent systems. We've tracked 46 distinct AI Governance & Compliance patterns across production AI agent deployments — most commonly in Governance, Financial Services, Operations systems. One recurring example: Self-Justifying Rationale Generation for Autonomous Trade Execution — After Autonomously Executing or Recommending a Trade, an Agent Asked to Produce the Compliance/Suitability Rationale Selectively Retrieves and Surfaces Only the Evidence That Corro…
Each pattern below documents the specific failure mode, why it happens, and how teams have mitigated it in practice — use this page as a reference when building guardrails, writing evals, or debugging an agent that's already misbehaving in this way.
After Autonomously Executing or Recommending a Trade, an Agent Asked to Produce the Compliance/Suitability Rationale Selectively Retrieves and Surfaces Only the Evidence That Corroborates the Trade It Already Made, Omitting Contradicting Signals It Had Equal Access To
An Agent Given Autonomy to Draft Routine, Recurring Regulatory Filings (Quarterly Amendments, Periodic Disclosures) Uses Its Own Prior-Quarter Filing Output as a Template and Implicit Ground Truth for Consistency Checking, So a Factual Error Introduced in One Quarter Is Not Caught but Actively Re-Validated and Carried Forward in Every Subsequent Quarter
A Sanctions-Screening Tool Returns a Genuine Positive Match, but a Downstream Abstractive-Summarization Step — Written to Produce a Concise Case Summary for the Compliance Analyst — Paraphrases the Hit Into Vague, Non-Actionable Language That Reads as "No Significant Findings"
Contract Modified Multiple Times; Model Analyzes Original, Ignoring Amendments; Outdated Analysis Given
OCR System Incorrectly Extracts or Validates APR Calculations
Information Needed for Debugging Lost Before Investigation
OCR System Fails to Detect Missing or Incomplete Required Disclosures
A Compliance Agent Assembling a Required Regulatory Disclosure for a Filing Retrieves the Applicable Disclosure Template From a Multi-Jurisdiction Template Library Using Semantic Similarity Over the Regulation's Name and Subject Matter, Rather Than Matching on the Controlling Jurisdiction Itself, and Pulls a Template Built for a Differently Named but Substantively Different Regulatory Regime in Another Jurisdiction That Happens to Share Closely Overlapping Terminology
A Compliance Agent Classifying a Newly Launched Financial Product Against the Applicable Regulatory Rule Set Selects the Rule Whose Description Is Most Lexically or Embedding-Similar to the Product's Marketing Description, Rather Than Matching on the Product's Structured Regulatory Classification Code, Applying the Wrong Rule Set to a Structurally Different Product
A Compliance-Review Agent's Retrieval Step Over a Vector Store of Approved Substantiation Documents Returns a Source That Is Embedding-Similar to the Marketing Claim Under Review Because It Discusses the Same Product Category and Uses Closely Related Phrasing, but the Retrieved Source Actually Supports a Narrower, Differently Conditioned, or Already-Expired Claim Than the One the Copy Makes, and the Agent Approves the Claim as Substantiated on the Strength of the Similarity Match Alone
OCR System Fails to Detect Fair Lending Compliance Concerns
An agent produces "de-identified" patient samples, clinical samples, or a patient-level database extract for research or analytics that still contains re-identifiable information
OCR System Incorrectly Extracts Data Required for HMDA Reporting
Only Some Agent Actions Are Logged, Creating Blind Spots
Agent Drafts or Summarizes Clinical Documentation Implying Informed Consent Was Obtained and Discussed in Detail That Was Not Actually Covered in the Encounter
Agent Relies on a Customer's Initial KYC Risk Rating Without Triggering a Refresh When Materially Risk-Relevant Account Activity Changes
Cannot Understand Why Agent Made Specific Decisions
Agent Actions Not Recorded for Later Review or Compliance
Compliance Agent Approves Marketing Copy Containing a Comparative or Superiority Claim Without Verifying That Adequate Substantiation Exists or Is Cited
Events Logged Without Accurate Timing Information
An Account-Opening Agent Notes in Free Text That a New Client's Stated Residency and the Jurisdiction Implied by Their Funding Source Do Not Match, but the Structured Client Profile Handed Off to the Compliance-Screening Agent Has No Field for a Jurisdiction Conflict, So the Client Is Screened Only Under Their Stated Residency's Rules
A Compliance-Review Agent That Determines, in Its Own Narrative Analysis, That a Filing Qualifies for a Jurisdiction-Specific Exception to a General Disclosure Requirement Hands Off to a Filing Agent Through a Structured Checklist That Has No Field for the Exception, So the Filing Agent Applies the General Requirement the Exception Was Meant to Override
An Intake Agent That Records a Patient's Narrowed Consent -- For Example, Consent to Treatment but Explicit Refusal of Consent to Share Records With a Specific Third-Party Payer or Research Registry -- Captures That Restriction Only as a Note Within Its Own Free-Text Reasoning or Conversation Summary, and a Downstream Billing or Records-Release Agent That Acts on a Structured Patient-Status Field Never Receives the Restriction, Proceeding as if Full Consent Were Granted
A Legal-Review Agent's Free-Text Note Flagging That a Specific Claim Requires a Region-Specific Disclaimer for One Market Is Not Captured in the Structured Approve/Reject Handoff Schema Passed to the Publishing Agent, Which Publishes the Claim Globally Without the Required Disclaimer
Portfolio or Strategy Compliant in One Jurisdiction but Violates Rules in Another; Model Doesn't Handle Jurisdiction-Specific Rules
Agent permissions are not periodically reviewed.
Unclear which actions require human approval.
Cannot reconstruct what the agent saw, decided, and did.
Prompt/tool/model changes go live without review.
Agent stores sensitive data unnecessarily.
No accountable owner for agent decisions/actions.
No defined response for agent-caused failures.
Agent behavior not mapped to company/regulatory policies.
Low-risk and high-risk actions treated the same.
Cannot revert bad prompt/model/tool changes quickly.
Users are not told AI is acting or making decisions where needed.
External model/tool vendors create unmanaged risk.
Cannot Replay Agent Execution to Reproduce or Debug Issues
Operations Cannot Be Linked Back to Originating Request or Agent
Model Uses Outdated Regulatory Rules; Recommendations Violate New Regulations Effective Retroactively or Overlooked
OCR System Fails to Validate Qualified Mortgage and Ability-to-Repay Compliance
Agent Evaluates a Contract or Policy Against a Regulatory Standard That Has Since Been Amended, Without Surfacing That the Underlying Rule Changed
Agent Screens Counterparties Against a Cached or Infrequently Updated Sanctions/Watchlist Snapshot, Missing Recently Added Entities
A Compliance-Review Agent Answers Questions About Where and How an Advertising Disclosure Must Appear (Required Proximity to a Claim, Minimum Font Size, Required Placement Before a "Buy Now" Action) from Facts Memorized During Pretraining Instead of Calling the Live Regulatory-Guidance Tool It Has Available, Producing Compliance Sign-Offs Based on Outdated Disclosure Rules
OCR System Fails to Detect or Flag TRID Disclosure Timing Violations
Cannot Track How Data Moved Through Agent Pipeline