Who writes this

Your Name. Engineer writing about how to deploy AI agents safely and keep them under control.

[Replace with two or three sentences of background: what you have built or run, and why that gives you a useful view on agent governance.]

Why this exists

Agents now read email, call APIs, run code and move money. Most writing about them is either marketing or abstract principle. The people who actually have to decide what an agent may do (the engineer wiring up the tools, the product manager scoping the feature, the compliance lead signing off) need something in between: concrete patterns, honest comparisons and working code.

Each piece aims to answer one practical question: what control do I put where, and what does it cost me?

Editorial policy

  • Independent. No sponsored posts, no affiliate links, no paid placements. If I have a relationship with a company or project I write about, I say so in the piece.
  • Vendor-neutral. Tools are compared on published documentation and hands-on testing, with the method stated. Open-source and commercial options are treated the same way.
  • Sources cited. Every factual claim that is not common knowledge links to a primary source where one exists: the regulation text, the standard, the paper, the docs. Each article ends with a sources list.
  • Corrections noted. Errors are fixed in place, and a dated note at the end of the piece says what changed. Substantial revisions update the “Updated” date at the top.
  • Not legal advice. Pieces on regulation explain what the texts say and how teams interpret them. They are not a substitute for advice on your situation.

Spotted an error?

Reply to any newsletter email. Corrections are always welcome and always credited if you would like.