Prototyping & Greenfield
Get comfortable with AI agents
Start with a new project. Your team builds a prototype and practices working with agents without touching the existing codebase.
- Get hands-on with Cursor, Claude Code or Copilot.
- Turn an idea into a specification and prototype.
- Work in pairs and share your results in a demo.
Your Codebase, Your Backlog
AI agents on your real tasks
Bring your own codebase and backlog. Your team works through real tasks with agents, from understanding the code to reviewing the result.
- Set agents up with context for your codebase.
- Research → Plan → Implement on real backlog tasks.
- Demo the work, review results and agree next steps.
Harness Engineering
From individual productivity to team capability
Build and test a component of your shared agent environment. Explore how instructions, tools and feedback make agents more reliable across the team.
- Explore the harness across your delivery workflow.
- Build a harness component or shared team tool.
- Share results, learning and improvements to try next.
Team Enablement
A shared plan for better ways of working
Take stock of how your team works with agents. Choose the improvements that matter and leave with clear owners and a first experiment.
- Map current practice and delivery bottlenecks.
- Choose priorities and accountable owners.
- Build a prioritised roadmap for your team’s harness components.