Could You Reconstruct What Your AI Did Six Months Ago?
Examiners ask what an AI system was asked and answered on a past date. Many teams expect their AI provider to hold that record. Usually it doesn't.
Blog
Production AI, governed.
Analysis of AI regulation and examination practice for financial services, insurance, healthcare, and HR.
AI Guardrails at Financial Institutions: Can You Show the Control Ran?
Many institutions write AI guardrails as system prompt instructions. Why that falls short in an exam, and what a control an examiner can rely on looks like.
Making AI Audit Evidence Independently Verifiable
How hash chains, independent timestamps and offline verification let a reviewer check an AI audit trail for themselves.
The Global AI Law Landscape
How 24 jurisdictions across the Americas, Europe and Asia-Pacific regulate, guide and promote AI, and the common rules they share.
Seven Financial Services AI Rule Changes from Q3 2026
Third-party risk, the NAIC bulletin and evaluation tool, FINRA Rule 2210, Regulation B and the EU Omnibus: the AI rule changes for the CCO's Q3 review.
SR 26-2 Has No Checklist. How to Find the Gaps in Your Model Risk Program
SR 26-2 sets principles for model risk. Where programs fall short, how examiners test each area, and a free eight-area gap assessment.
The NAIC AI Systems Evaluation Tool: What a Chief Compliance Officer Should Have Ready
The NAIC's AI questionnaire for insurers is planned for adoption in November 2026. What it asks, and the records a compliance officer should have ready.
Meilynx at IAPP AI Governance Global 2026 in Seattle
We're attending IAPP Privacy. Security. Risk. + AI Governance Global in Seattle, October 6-9. Here's where to find the Meilynx team.
Meilynx at the ISG AI Impact Summit in New York
We're pitching in the ISG Startup Challenge at the AI Impact Summit in New York on October 28-29, with the latest Meilynx features for regulated AI.
Catching AI Agent Drift Without Exposing Your Prompts
AI agents can change after approval without a new release. How permission and prompt drift happen, and how to catch them while prompts stay in your environment.
How a Proxy Governs MCP Agents
What runtime governance requires at the MCP layer, and how a proxy between agents and their tools provides it.
We Put Our Proxy in Front of Professional Attackers
Meilynx completed an independent, authenticated penetration test of the proxy in August 2026. What was tested, what was found, and how to read the report.