AI market research can mean two very different things: a model that guesses what customers think, or a model that talks to customers and reports what they said. Moderators does the second. AI moderators interview your real customers in chat or voice, and every finding links back to the quotes that support it.
Try to export last month's report. Say what you are thinking.
I expected it under Reports, so I am looking there.
What made you look there first?
The menu says Reports, and I assumed exports live there.
Quote from the interview
“Nothing told me if I had done it right.”
It can be, when the tool invents respondents or answers on their behalf. Moderators takes the opposite approach. Every answer comes from a person you invited, and the AI's job is to ask well, listen and organise. That distinction matters because a finding you cannot trace to a human sentence is an opinion, not research.
In three places. It drafts the interview guide from your goal. It conducts each interview, deciding on the spot which answer deserves a follow-up. And it analyses all transcripts together, grouping what people said into themes ranked by how often and how strongly they appear. The people, the answers and the quotes stay human.
The slow parts of qualitative research are scheduling, moderating and coding transcripts. Here the first two disappear because participants answer through a link when they choose, and the third is done as interviews finish. A study that would normally need weeks of coordination can be read as results arrive. How the pieces fit together is described in how AI-moderated research works.
It is built to dig. If someone says a tool is too expensive, the moderator asks what they compared it with and what they pay today. If someone says they switched, it asks what happened the week before. You can read each exchange in the transcript and judge for yourself, which is the right way to evaluate any AI moderator. For a closer look at the interview style, see AI-moderated interviews.
Moderators does not supply respondents, predict demand from nothing or produce statistics about a market at large. It reports what the people you invite tell it. If you invite a list of customers, you learn about those customers, in depth.
Each of these is a single sentence in the "What do you want to learn?" field, and each produces a guide you can edit before sending.
Annual billing is the default and halves the monthly price. There is no free plan, and limits spent in a period do not return.
On Starter, $99 divided by 100 chat interviews is about 99 cents per interview at the limit, and less on annual billing. Full detail is on the pricing page.
Enter your question, answer as a respondent and read the first theme with its quote. Your data stays yours.