
Choose a real task, build an agent to help, and learn to improve it yourself. Use the roadmap below to find your next skill and put it into practice.
Get your AI workspace ready.
Follow the setup guide for Claude or ChatGPT, with screenshots and simple checks to confirm you’re ready.
Open the workspace prep guide
↖ Hover, tap, or focus any segment to see what it does.
Collaborative framing
Open the problem as a possibility. This invites exploration before the team locks into the first solution.
Action
Name the exact work to perform: rewrite, compare, summarize, prioritize, design, or build.
Definition of done
Describe the observable result you want, so the agent can tell what success looks like.
Boundary signal
Introduce the tradeoff you refuse to make. This turns a preference into a clear guardrail.
Rules + boundaries
State what must not change, what needs approval, and any limits on scope, time, tone, or risk.
Source cue
Tell the agent to ground its answer in the material that follows instead of relying on default assumptions.
Sources
Name the files, systems, examples, research, or data the agent should use to do the work.
USE THE SMALLEST GEAR THAT CAN OWN THE RESULT
Sets direction · defines good · reviews the work · owns the outcome
Rate your current usage and confidence in each capability. Your red and yellow ratings shape the AI Roadmap below.
Use the same Red, Yellow, and Green scale. A red or yellow rating adds a coaching call to your AI Roadmap because matching self-guided curriculum is not available yet. Related foundation modules may introduce relevant concepts, but do not teach the complete advanced control. These optional ratings do not change your core 12-capability score.
Email coaches@technical-leaders.com and tell us where you’re stuck: choosing the gear, closing a capability gap, or reviewing a workflow before you trust it. A coach will help you think through the next move.