Wisp replaces the static questionnaire with an AI assistant that interviews every respondent — by voice or chat — probes the answers worth probing, and reads how they feel. You get clean, structured data and the story behind it, at the scale of a form.
Event Registration Intake
A friendly conversational intake for an event sign-up.
StartProduct Feedback Conversation
A short conversational survey that probes the 'why' behind a product rating.
StartThe Vibe Check ✨
A 5-minute chat with Max that reads your vibe — no boring questions, just genuinely good conversation.
StartWisp's conversation engine decides every turn — acknowledge, dig deeper, steer back, or move on. One conversation, captured twice: a clean value and the words behind it.
Most tools bolt an AI button onto a static form. Wisp is built the other way around — an agent authors the form, an agent conducts it, and an agent helps you make sense of the answers.
Describe what you want to learn and Iris drafts it. The Form Architect builds and rewrites forms by voice, chat, or an uploaded brief — PDF, DOCX, CSV — with a live diff and one-click undo.
Every response is a real conversation. The Assistant asks non-linearly, probes the “why”, adapts its tone to how people feel, and works by voice or chat.
Ask your results questions in plain English. The Insight Analyst answers with charts, cross-tabs, and real verbatim quotes — every number grounded in deterministic compute, never invented.
A normalized value for every question — plus the respondent's own words. Chart it and quote it.
Acknowledges, probes interesting answers, and adapts non-linearly — and it's emotionally aware, leading with empathy when someone's frustrated instead of a cheery “sounds great!”.
Speak to the form and it speaks back. Any answer is speakable; controls are an optional tap.
Ask your data questions in plain English. Iris answers with charts, cross-tabs, and verbatim quotes — every figure grounded in real compute, not guessed.
The Form Architect drafts and improves forms by chat, with / commands, # references and @ mentions.
Keys stay server-side. The Assistant always discloses it's AI. Insights cite real evidence.
A natural interview. Probes the interesting answers. The default.
A cascading pipeline — speech-to-text, the conversation engine, text-to-speech — lets respondents answer out loud, and hear the Assistant reply in a voice you choose. It falls back to chat seamlessly, even to the browser's own speech with no keys.
A Wisp form is conducted by an AI assistant that interviews each respondent in a natural conversation by voice or chat, asks follow-up questions on interesting answers, and captures both a clean structured value and the narrative behind it — instead of a static list of fields.
Yes. Wisp uses a cascading voice pipeline (speech-to-text → conversation engine → text-to-speech) so respondents can answer out loud and hear the assistant reply, with live captions and a seamless fallback to chat.
Traditional tools add AI as a feature on top of static fields. Wisp is AI-first: an agent authors the form, an agent interviews every respondent and probes the “why”, and an agent helps you analyze the answers — giving you the depth of an interview at the scale of a survey.
No. Wisp runs end-to-end with no keys using a built-in mock model and the browser's own speech, so you can try a full conversational form immediately; adding a model key upgrades it to a production LLM and premium voices.