Qualitative evidence is rich, but slow to analyse, hard to compare and easy to distort. Coding one 90-minute group by hand takes days, summaries over-weight whoever spoke most, and generic AI summarisers can paraphrase quotes and invent consensus.
Room Wisdom takes the transcript of a focus group or interview, in any of the 24 official EU languages, and maps who said what and who took which position. Each participant is recorded as for, against, mixed or silent on every proposition, and silence is never counted as agreement. A researcher approves the key judgement calls, and every quote is checked word for word against the transcript before anything is published. Names are removed before analysis.
In a first pilot, one 30-minute focus group held in Greek with six participants, all 188 quotes were verified word for word against the transcript, coding consistency reached κ 0.95 (κ 0.93 on stance judgements), and processing took about 120 minutes. These are results from a single pilot session, not general accuracy rates, and a benchmark against human coding is still running.
We are looking for partners to pilot Room Wisdom on live studies of 4 to 20 groups or interviews, to write it into EU, national or client proposals, or to join as launch partners. Get in touch at info@ethoslab.gr.