Skip to content
Pondara Pondara docs

AI as sparring partner

The most-asked question about Pondara is: “What does the AI actually do here?” The short answer is that the AI is a sparring partner, not a tutor. It plays the people you encounter in a scenario, and it opens reflective questions after the action. It does not transmit content. It does not grade you. It does not replace your trainer.

What the AI does

Three things, on three different surfaces.

  • Plays the actor in the hybrid dialogue. When a scenario calls for a customer, a patient, a charge nurse, a workshop foreman, the AI is in role as that person. It speaks from the actor’s brief — what they want, what state they are in, what they will and will not say. It does not break character to coach you.
  • Proposes evidence after the engagement. Once the dialogue closes, the AI extracts what it noticed and proposes a signal for each affected sub-competency. That signal is visible to you and to your trainer, but it does not move your record on its own.
  • Opens reflective prompts. During the Reflect phase, the sparring partner asks open questions about what you did and why. These are prompts, not answers. The AI does not tell you whether you handled the situation well.

What the AI does not do

This list earns trust by being precise.

  • It does not tell you the right answer. Even if you ask. Most professional situations do not have one. The AI’s job is to make you face the situation, not to spare you from it.
  • It does not replace your trainer’s verdict. The AI proposes; the trainer decides. Your competency record only moves on a trainer verdict.
  • It does not act outside its in-scenario role. During Attempt the AI is the actor — it does not pop out of character to summarise, hint, or assess.
  • It does not see your record across organisations. What happens in your org stays in your org. The AI you talk to in a scenario has the scenario brief and the dialogue so far. It does not have a cross-org view of you.
  • It does not train on your engagements. Your dialogue is yours and your org’s. It is not used to train future models.

AI opt-out

You can disable AI-authored evidence in your settings. When you do, the AI no longer proposes signals on your engagements. Your trainer authors evidence directly from the dialogue and writes the verdict themselves.

The AI sparring partner inside the scenario is unaffected. The actor still plays its role; the dialogue still works the way it always did. What changes is what happens after: the AI does not write a first read, and your trainer does the extraction by hand.

Some learners prefer this — the AI signal can feel like a verdict even when it is not. Some trainers prefer it for high-stakes scenarios. The toggle is yours.

See How to opt out of AI evidence for the step-by-step.

Where to look next