Review a roleplay meaningfully
Review the conversation before the score. Identify what the client asked, what you answered, which claim needs evidence and one behaviour to practise again.
Book a WhatsApp demoRead four pieces of evidence
- What exact concern did the client express?
- What question did the adviser ask to clarify it?
- Which statement was factual and what supports it?
- Was the proposed next step optional and proportionate?
Write one change
A note such as “be more confident” is difficult to act on. Replace it with an observable change: “Ask which part of the payment commitment concerns the client before explaining options.”
Repeat and compare
Use the same fictional scenario again. Compare whether the missing question was asked and whether the response addressed the concern. Preserve any remaining uncertainty instead of marking the whole exercise as passed.
Fictional review note
Observed: the adviser explained benefits before clarifying cost. Next exercise: acknowledge the concern, ask whether it relates to amount or duration, then pause for the answer.
What to keep in mind
- AI scoring is feedback, not a professional qualification or proof that a statement complies with every applicable requirement.
Questions, answered
How many changes should I make at once?
Choose one observable change so you can tell whether the repeat exercise improved it.
Can a manager use only the AI score?
A reviewer should inspect the underlying conversation and independently check factual claims.
Sources
- Insurance Authority — Code of Conduct for Licensed Insurance Agents Source version date needs checking
- PCPD — Checklist on Guidelines for the Use of Generative AI by Employees Source publication month
Check each original document for its date, definitions and scope. A citation does not confirm the latest version or imply endorsement of our examples.
Editorial policy