Would You Let an AI Agent Move Your Money?

What human-in-the-loop costs once it stops being a stub

Part 4 of an ongoing experiment: building an LLM-powered support agent with deterministic boundaries. The companion repo grows with the series.

The rules say the customer is owed a €512.64 refund. The agent agrees. The API is one method call away. Who presses go?

That one line of code is where "AI-assisted" becomes "AI has authority." An agent can be perfectly capable of deciding that a refund is justified without being allowed to issue the refund. Deciding and doing are different permissions. That's the boundary I wanted to make impossible to blur.

The decision: separate judgement from authority (ADR 002)

Would You Let an AI Agent Move Your Money?

Other newsrooms on this story

Related reading

Let your LLM take real-world actions — without giving it the last word

A practical control checklist before an AI agent moves money

Your AI is about to start spending your money. Who checks its ID?

The hard part of an AI feature is knowing where NOT to use AI

Who pays when the AI hires another AI?

Human-in-the-Loop AI: Why 'Ask the LLM to Confirm' Isn't Enough