Engagement phases

An engagement is one learner running one scenario once. Inside a single engagement, you move through phases in order. The default sequence is the one most scenarios use, and it is the one this page describes.

Interpret

Before you act, you read the situation. What is happening here? What matters? What is missing? Who is in the room and what do they want?

There is no right answer in Interpret. You are not being scored. You are practising the habit of looking before doing — the reflex an expert in your field has and a novice does not. Your interpretation is recorded so you and your trainer can come back to it.

Attempt

Now you act. The scenario opens into the hybrid dialogue: you write what you say or do, and the AI replies in role as the actor. The AI is the customer, the patient, the supervisor — whoever the scenario calls for. It is not your guide. It will not tell you what to do next.

Your trainer can drop into the dialogue at any turn. When they do, the AI yields. This is not a simulation you run alone; it is a shared thread.

Reflect

After the action ends, the AI sparring partner opens reflective questions. What did you choose, and why? What did you notice that you did not act on? What would you do differently next time?

These prompts are open-ended on purpose. The aim is not to assess your performance — your trainer does that — but to surface the reasoning behind what you did. The answers feed your record alongside the dialogue itself.

Reflection is not assessment

Your self-rating in Reflect does not move your competency vector. Only your trainer’s verdict does that. See evidence and verdicts.

Theory support

Theory support is an optional phase. It does not open automatically. It surfaces only when the AI’s reflection turns suggest a specific gap — “you mentioned you weren’t sure about the escalation rule; here’s the section of the protocol that covers it” — or when you ask for it.

The idea is the inversion of an LMS: theory arrives because action exposed a need for it, not the other way around. Reading without a question to answer is forgettable; reading after the question landed is not.

Re-attempt

Some scenarios are worth a second pass. After Reflect (and optional Theory support), you can re-attempt: run the scenario again with the same brief and the same actors, knowing what you know now.

A re-attempt is linked to the previous engagement in your history. Your trainer sees the pair side by side. The point is not to “get it right this time” — it is to give the reflection something to land on in your hands and your mouth.

Phases are configurable

The default sequence works for most scenarios, but a field expert can drop or reorder phases when the scenario calls for it. A one-shot crisis scenario might skip Re-attempt; a high-stakes procedure might require Theory support.

Where to look next

• AI as sparring partner — how the AI behaves differently in each phase.
• Evidence and verdicts — what happens to your record after the engagement ends.
• Scenario lifecycle — how the scenario you just ran reached you.
• Sub-competencies — the eight dimensions an engagement contributes evidence to.