AI agent for executive coachs
Leadership Behavior Experiment Agent
Turn each coaching goal into a series of small behavior experiments, each building on the result of the last.
What it does
Coaching goals such as listen more or delegate better stay abstract because there is no small test of the new behavior. For each goal, the agent designs a small experiment the client can run in a week, such as handing off one task in a meeting, and sets how it will be observed. It collects feedback from the client and, with permission, from peers, and compares it with the baseline from the first session. It then judges whether the experiment worked, partly worked or failed, and proposes the next one. If feedback is thin, it asks again. The coach approves each experiment and any contact with peers. Edge case: the client reports success, but two peers saw no change, so the agent proposes a clearer experiment.
How it works
Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.
Read the steps as a list
- Coaching goal set or experiment period ends
- Read the goal, baseline and previous experiments
- Design a one week experiment with an observable sign
- Coach approves the experiment and any peer contactThe agent waits here for your OK.
- Collect the client's report and permitted peer feedback
- Compare feedback with the baseline
- Is there enough feedback to judge the result?If not: Ask the client or peers a short follow-up question. Back to step 5.
- Rate the result as worked, partly worked or failed
- Do client and peer views agree?If not: Design a clearer experiment that makes the behavior easier to see. Back to step 3.
- Coach approves the next experimentThe agent waits here for your OK.
- Experiment log and next step
How it decides
An experiment counts as worked when client and peer feedback both move beyond the baseline. A mixed result leads to a clearer experiment; a failure leads to a smaller or different one.
- Each experiment must be observable by another person
- At least 2 feedback sources before rating
- A failed experiment gets a smaller next step
- Never contact a peer without client and coach agreement
Make it yours
Every agent is a starting point. You choose these settings for your own situation.
- Experiment length (default 1 week)
- Feedback sources required
- Rating scale
- Who may be contacted
What keeps you in control
It always asks you first
- Coach approves each experiment
- Coach approves any peer contact
Hard limits
- Never contacts peers without approval
- Never shares peer comments with the client unless the coach agrees
It stops when
- Done: goal reaches the target behavior
- Stop: client stops the experiments
Set it up
We guide you through the set-up, step by step
Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.
- One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
- The agent then walks you through connecting your own data, one source at a time
- A downloadable copy with the flow chart, the rules and the full guide
