Chris Conen/Writing/Proof over promises
Writing on agent accountability.
Problem-first essays on making AI agents and automated workflows provable: tamper-evident audit trails, verifiable logs, n8n, and the record-keeping rules now arriving with the EU AI Act.
-
The "AI Did It" Defense Is Not a Defense
If your AI books, prices or messages on your behalf, "the AI did it" won't get you off the hook. Here's what actually protects you when a decision is questioned.
-
Your AI Made a Mistake. Can You Prove It Wasn't Your Fault?
When an automated workflow goes wrong, the fight is usually about whose fault it was. Here's how to find out in seconds — and prove it — instead of guessing.
-
Three Questions to Ask Before Letting AI Touch Your Customers
Before you let AI book, price or message your customers, three plain-English questions decide whether you'll be able to stand behind what it does.
-
How a Signed Audit Layer Exposed Four Silent Bugs in a Live Workflow
A first-person narrative on how adding a cryptographic, verifiable audit layer to an n8n booking workflow exposed four active, silent system bugs.
-
How to Add a Tamper-Evident Audit Trail to n8n (Step by Step)
A step-by-step tutorial: add an independent, cryptographically signed audit trail to any n8n workflow with one Code node, verifiable offline from a public key.
-
EU AI Act Article 12: Demystifying Event Logging for AI Agents
An explainer on EU AI Act Article 12 logging requirements for AI agents, Article 26 deployer obligations, and how to address the GDPR Article 17 tension.
-
Storytelling vs. Evidence: Why Agent Logs Aren't Evidence
A conceptual exploration of why traditional application logs aren't evidence, what a tamper-evident audit trail for AI agents requires, and why signatures matter.
-
How to Verify an Agent Log Offline with Only a Public Key
A complete step-by-step technical guide on how to verify agent log offline using the open-source AXR CLI and standard Ed25519 cryptography.