Jul 6, 2026
An AI governance framework you can actually run
NIST AI RMF and ISO 42001 say what good AI governance looks like. Here is how to run it for agents that act — mapped to controls a platform can enforce.
ReadTag
11 posts tagged “Agents”.
NIST AI RMF and ISO 42001 say what good AI governance looks like. Here is how to run it for agents that act — mapped to controls a platform can enforce.
ReadDescribe what you want and get an agent — every 2026 tool sells this. The question that matters: can a no-code agent act safely, under review, on real data?
ReadAn AI notetaker whose summary nobody acts on is theater. The real test: whether decisions and action items become tracked, owned, governed work — not a doc.
ReadAI can draft the sprint in seconds, but the plan is still yours. How to use an agent to prep, scope, and forecast a sprint without losing the judgment calls.
ReadAs AI agents move from suggesting to acting inside your work platform, 'who approved this?' stops being a nice-to-have and becomes a compliance question. A practical look at why propose-and-approve, immutable audit trails, and per-workspace cost control are becoming table stakes.
ReadNo — but the job is changing fast. AI is on track to automate most routine PMO work by 2030; what it can't do is own the consequential decisions. An honest look at which parts of the PM role vanish, which get harder, and what stays human.
ReadEvery work tool now has AI. The model isn't the moat — anyone can call the same API. The durable advantage is governance: who can prove their agents are safe to turn on.
ReadFull autonomy is reckless and full manual review defeats the point. Propose-and-approve is the middle path — and getting the boundary right is the whole design problem.
ReadFully autonomous agents are a demo; supervised agents are a tool. Why the propose-approve boundary — not raw capability — is what makes AI agents safe to leave running on your real backlog.
ReadThe Model Context Protocol lets any AI client reach your work tool's data and actions through one standard interface. Here's what it is, why it matters, and how to expose it without handing an agent the keys.
ReadA single all-powerful AI is harder to trust, harder to bound, and harder to debug than a team of small, scoped agents. The design pattern that makes multi-agent systems both safer and more capable.
ReadTry it