Capability Conversation – Start Here

START HERE: Run the Capability Conversation Diagnostic

Facilitator guide for the 26-page quick scan, break synthesis, and follow-up investigation

Purpose: Use the capability pages to learn what is happening across the system, not to test framework compliance or assign a maturity grade. The first pass gathers evidence and scores. A deliberate break supports cross-capability synthesis. The second pass investigates the most useful patterns and chooses a bounded next step.

Before the conversation

  • Agree on the system boundary, purpose, and time available.
  • Interview the sponsor or contact first; avoid surprising people with known sensitive facts.
  • Invite a cross-section of roles, functions, products, and decision levels.
  • Decide how notes and differing perspectives will be protected and shared.
  • Prepare one scoring sheet that can preserve ranges, evidence, and confidence.

Pass 1: quick scan

  1. Show one capability page at a time and briefly orient participants to its meaning.
  2. Ask what actually happens.Use a recent example rather than the intended process. 
  3. Have people choose 1, 2, 3, 4, or U. Record the range before discussing it.
  4. Capture evidence, consequence, who is affected, perspective differences, and confidence.
  5. Move on. Do not diagnose causes or recommend a practice during the scan.

At the break: synthesize

  1. TellAmpGPT that you are the facilitator and have just completed a Capability Conversation diagnostic. 
  2. Provide the full scoring sheet, evidence, system boundary, participants, and important constraints.
  3. Request a short Facilitator Break Brief first: overall picture, strongest signals, clusters, disagreements, unknowns, competing hypotheses, and suggested re-entry.
  4. Then request the fuller analysis only where it will help the next conversation.

Suggested opening

“We will first walk through the whole system and record what we observe. We will not solve each issue as it appears. After the scan, we will look for patterns, consider more than one explanation, and decide what evidence or small action would be useful.” 

Scoring discipline

  1. 1 = material issue to investigate; 4 = effective for the current need; U = insufficient evidence.

  2. Score the capability, not the people.

  3. Preserve disagreement; do not average it away.

  4. A score is a navigation signal, not a cause, solution, or total maturity measure.

Facilitator guardrails

  • TreatAmpGPT relationships as hypotheses to review, not established causes. 
  • Keep credible competing explanations visible.
  • Do not expose private comments or infer motives.
  • A weak capability may be an effect, a local adaptation, or outside the chosen boundary.

Pass 2: investigate and act

  1. Resume with the brief and ask whether the pattern matches participants’ experience.
  2. Select one to three clusters; do not treat every low score as a separate project.
  3. Trace recent work and ask what evidence would distinguish thecandidate explanations. 
  4. Choose a small adaptive experiment with expected evidence, review date, adaptation signals, and possible harmful effects.
  5. Keep the full scoring set as a baseline for later comparison, not as a target dashboard.

Copy into AmpGPT at the break

I am facilitating a Capability Conversation and have just completed the first-pass diagnostic. 
System boundary: [boundary]. Participants: [roles]. Context and constraints: [context]. 
Below are the scores, score ranges, evidence, consequences, confidence, and disagreements for all 26 capabilities. 
 
First give me a concise Facilitator Break Brief covering: 
• overall picture and strongest signals; 
• cross-capability clusters, disagreements, and unknowns; 
• two or three credible hypotheses and evidence that would support or challenge them; 
• the most useful conversation to resume with. 
 
Do not treat scores as causes or prescribe a practice yet. Then provide the fuller analysis and conditional next-step options.