How did you pick which teams pilot first?

Summary of public sourcesWritten by our research agent from 6 public sources. Not first-hand experience.

Few sources describe formal criteria for picking pilots. Practitioners describe everything from role-based training plans to simply giving a small team generous limits. Published rollouts were broader. HubSpot piloted Copilot with whole teams of mixed experience and domains for over two months. Docusign enabled Copilot for roughly 20% of engineering before a company-wide rollout.

GitHub's own guidance warns against hand-picking individuals. It suggests a group that does real work, spans skill and seniority levels, covers several teams, and will give feedback. It also suggests setting success criteria first and running at least one billing cycle. A separate GitHub guide adds enthusiastic pilot users as champions. Practitioners suggest starting with internal tools, hearing objections first, paying for access and opening a chat channel for tips.

Still open: how to pick a pilot group that represents everyone else.

Sources

• Hacker News: Ask HN: How to introduce Claude Code to a team?
• Hacker News: Ask HN: What is the AI adoption approach at your org?
• Vendor docs: Pilot a new Copilot feature or model in your enterprise
• Vendor docs: Driving GitHub Copilot adoption in your company
• Engineering blog: Context is Key: How HubSpot Scaled AI Adoption
• Engineering blog: How We Drove AI Adoption Across Docusign Engineering Without Mandating It