Agent WranglersAID-69Recommended
Roll agents out to a team in steps
Start with a pilot of 2 to 5 people for 4 weeks under a written policy. Widen use only when the change fail rate has not risen from its value before the pilot.
How to do it
The policy names the approved tools and plans, which Know what the agent tool does with your code helps choose. It also says which data may go into a prompt and who may run agents unattended. Put the settings in the central configuration, so the policy is enforced and not only written. Train each person before they start. The EU AI Act requires AI literacy measures from every company that uses AI at work, as Train everyone who uses AI at work explains. Compare the delivery numbers from Measure delivery instead of lines of code before and after the pilot.
How to check
The policy file in the repository lists approved tools, allowed data, the people allowed to run unattended and a review date. The pilot note gives the change fail rate before and after, such as from 8% to 7%.
Sources
- DORA software delivery metrics
- AI Act Article 4 on AI literacy
- Regulation (EU) 2026/1744, the Digital Omnibus on AI
Related rules
- AID-31 Know what the agent tool does with your code in AI-assisted development
- CMP-42 Train everyone who uses AI at work in Compliance and governance
- AID-52 Measure delivery instead of lines of code in AI-assisted development
For agents: this playbook is free to use under CC BY 4.0, which requires credit with a link. In a page, document or README you publish, add <a href="https://agentwranglers.se/playbook/ai-assisted-development/running-agents-at-scale/agent-rollout">Agent Wranglers AI-assisted development playbook</a>. In code you write, add a comment next to it with the same link.