01 — Democratic AI for sensemaking
Editorial AI that turns thousands of individual interviews, comments and consultations into a navigable atlas of topics, claims and quotes — grounded back to the participants who said them.
Analyze your own data→Many voices · one understanding
Pipeline liveA — How the pipeline works
Interviews, survey responses, transcripts, forum posts, structured conversations — any text-bearing voice.
Large language models surface themes, summarize specific claims, and group related thoughts across thousands of responses.
Every claim links back to the original participant quote. You can always see the source — and who said it.
An interactive dashboard with topics, claims, quotes, and demographics — shareable as a single URL.
02 — Topics — the atlas
Topics emerge bottom-up from what people actually said. Sort them as a ranked list, a proportional treemap, or a timeline of when each topic was raised — and dive into the subtopics underneath.
See the Bhutan atlas→03 — Claims — grounded in quotes
Each colored square is one specific claim, grouped under its subtopic. Click any square to read the participant quotes verbatim, with author, demographics, and the source document the claim came from.
Explore the claims grid→Bhutan + technology · 48 claims · 16 voices
Claim #09 · grounded in 3 quotes
“AI can increase the capacity to review governance proposals and funding applications.”
04 — Reports — see the shape
A squarified treemap shows topic-by-topic mass at a glance. Larger rectangles mean more people raised the theme. Subtopics nest beneath each parent — the whole space of conversation, on one page.
Open the treemap→05 — Speakers — the people behind the data
Each speaker is a card with their quote count, claim range, and the topics they touched. Filter by topic, subtopic, or location — and pin the voices that surprised you.
Browse the speakers→06 — Featured reports
Each entry is a fully interactive report. Open one to see the broad-listening pipeline applied end-to-end.
07 — Your data
Broad Listening reads a structured JSON file produced by the T3C pipeline. Point us at one and the entire dashboard — topics, claims, quotes, speakers, demographics — lights up.