Topic hub

AI Loops: everything in one place

An AI loop is a repeatable piece of work with a clear definition of done, delegated to an AI agent that keeps working and checking until that definition is met. The agent runs the loop. A human owns its specification and is called for exceptions, missing decisions or protected boundaries, not routine completion.

This page organises everything we've published on AI loops by what you need, rather than by publication date.

1. Understand AI loops

What is an AI loop?

The canonical definition, and why leaders should learn loops instead of agents.

2. Specify a loop

The six-question specification

The operating contract between the business and the agent, in six answers.

Learn the method →

3. See examples

12 worked AI loop examples

Meetings, sales, reporting, onboarding, support, finance, hiring, compliance, marketing and product.

Browse the library →

Understand AI loops

Specify a loop

Choose and build

See worked examples

Scale across the organisation

Avoid the failure modes