Running a live interview

Where to start

Open any persona’s detail page from the personas list and click the Interview tab. That’s where sessions start, live or completed. Each persona has its own session history; switching personas means switching contexts.

A turn, end-to-end

You type a question and hit Send message. Behind the scenes Candor does three things in sequence: it generates the persona’s response, runs a critic agent over that response for consistency, and then delivers it to you. You don’t see the intermediate generate/validate steps; you see the final turn with metadata attached.

Each delivered turn shows you:

• The answer text.
• Confidence — a percentage indicating how certain the persona is in this answer. Low confidence is a signal that the persona is reasoning beyond what their memory firmly establishes.
• Uncertainty flags — short tags noting where the persona’s answer leans on inference rather than direct evidence.
• Memory types used — which of the four memory banks (identity, behavioral, belief, language) the persona drew on. Useful for understanding why an answer came out the way it did.
• Critic flag — if the response triggered a hard contradiction with established beliefs, you’ll see a badge marking the turn. Soft tensions (like a mood shift) are caught silently.

What the critic catches

The critic is a separate agent. Its job is to make sure the persona doesn’t drift mid-conversation. There are two severities. Hard contradiction: the response directly conflicts with an established belief or prior turn. These are caught and either regenerated or flagged so you can judge. Soft tension: the response is consistent but the tone or emphasis shifted in a way worth noting. These are tracked in the background but not surfaced inline.

The critic is what stops a Candor interview from devolving into agreeable AI prose. A persona who said earlier price isn’t the issue can’t turn around three turns later and say price is everything without either explaining why their stance changed or being flagged.

Memory persists

This is the part most people miss. The persona remembers the entire conversation, and they remember every prior session you ran with them inside the same project. End a session, come back tomorrow, start a new one, and the persona will refer back to what you discussed last time, including stance changes (which get logged explicitly, with a reason).

You can see prior session summaries from the persona detail page. They’re generated automatically when a session ends, distilling what was discussed and how the persona’s beliefs moved.

Refreshing world context

Personas have a snapshot of relevant industry/region context from the time the audience was generated. If you’re running a session weeks later and want the persona to know about something that’s changed in the world, Refresh world context rebuilds that snapshot. Use it sparingly — it pulls fresh evidence and recalibrates.

Ending a session

Click End interview when you’re done. A confirmation appears. Once you confirm, Candor generates a session summary from the transcript and the session moves into the persona’s history. You can re-open the transcript anytime, but you can’t add to a closed session — start a new one if you have more to ask.

When to use Mode 1 vs Mode 2

• Mode 1 (live) is for exploration and follow-your-nose questions. You learn faster when you can react to a surprising answer.
• Mode 2 (auto) is for coverage. When you need every persona answering the same set of questions consistently, the auto interviewer is more thorough than you’ll be by hand.
• Many projects use both. Mode 1 to find what matters, then Mode 2 to cover it across the audience.

Where to go next

• Running an auto-interview
• Building an interview guide
• Reading a persona profile