Your AI interviewer can do far more than ask questions — give it a knowledge base and it becomes a real expert, able to answer respondent questions and resolve issues right inside the conversation. No more "I don't know, I'll pass this along." Your interviewer can help immediately.
Two layers of knowledge
- This form's knowledge — documents and links that belong to just this form. Your interviewer checks here first; it is the primary reference.
- Company knowledge — a reusable library you build once in your dashboard and attach to any form as background context (policies, product docs, help pages).
What you can add
- PDF, DOCX, CSV, TXT, or Markdown files — uploaded directly.
- Web page links — ChatWisp reads the page for you, and links can auto-refresh weekly (or on demand), so a page you point at stays current without re-uploading.
Control when it helps
- Answer questions — the interviewer helps when someone asks or raises a concern.
- Proactively — it may also volunteer a relevant fact without being asked.
- Off — knowledge stays private and the interview behaves normally.
Guide the helper
Write short instructions ("answer billing questions from the policy doc; keep it brief") and an escalation message for when your knowledge has no answer ("I'll pass that to our team — they'll follow up by email"). Your interviewer will never make up an answer. Respondents see a subtle "from {your source}" note when it answers from your material — building trust without breaking the conversational feel.
The closing concierge
Flip the "Help them on the way out" toggle and your closing moment becomes proactive: once the questions are done, the interviewer reviews what the person shared, points them to the right help from your knowledge, and asks if it can answer anything else — then keeps helping until they are satisfied. A great way to turn a survey into a genuinely useful moment.
Related
The research motive & guardrails · Conversation style & how deeply the AI probes