Your next campaign, launch or AI feature deserves better than a chatbot's average of the web. Kinmat answers your questions with what real people said in real interviews across 11 countries, about how AI is entering their money, health, food, work and relationships. Every answer shows who said it. None of it is in your competitor's ChatGPT.
Introductory access for marketing teams · $3,000 per quarter · whole team, no seats
Interviews by senior researchers from the UX Alliance network. 20 years of fieldwork, Fortune 500-tested
Ask any chatbot how people feel about AI and you'll get a confident answer in seconds. It's the same answer your competitors got, scraped from the same internet. Kinmat is what's not on the internet: 100% human conversations, recorded by researchers in the markets you sell to. Decide with something they can't copy.
ChatGPT, Gemini and Copilot summarize the web. So does your competitor's. Whatever edge it gives you, it gives them too. And sometimes it just makes the quote up.
"ChatGPT hallucinated and produced fictitious quotations when we tested it."UX researcher, on Reddit
A single agency study runs $25–65K, takes 6–12 weeks and covers one market. So research happens once a year. Campaign decisions happen every week.
"They wanted over $700K a year from us. We didn't renew."Insights lead, on a research platform
Panels give you percentages. They don't tell you why someone trusts an AI with their meals but not their money. Or who was actually on the other end.
"The results were from people 'pretending' to be our target audience."Research lead, on a panel tool
Every answer Kinmat gives you exists because a researcher sat down with a person, in their own country and language, and listened. That's the whole supply chain.
No synthetic users. No scraped forums. No AI-written quotes. If the conversations don't hold an answer, Kinmat tells you "we don't know that yet", and it goes on the list of things to ask next.
Illustrative photography. Participants' identities and recordings are never shown. That's part of the deal we make with them.
Kinmat starts with one deep, ongoing study, AI in daily life, run in 11 countries by researchers who live there. New conversations keep entering it, and more studies are being scoped. You ask it by theme, by country, by moment in someone's day.
Budgeting, meal planning, health questions, homework, work drafts. What gets delegated first, and what never does.
"What do people already trust AI to do for them?"
Money, kids, health, their face, their data. Where the line sits, and how it moves by country.
"Will they accept an AI in our checkout?"
What AI gave people that they now consider theirs. The wins your product can build on instead of promising.
"What should our AI feature actually promise?"
Where it fails them, embarrasses them or makes them feel watched. Said out loud, to a person, in their own words.
"Why did the last 'smart' feature flop?"
What people think AI is, who they think it works for, and how that changes what your message means to them.
"Does 'AI-powered' help or hurt in Germany?"
Where your brand, your competitors and your category show up unprompted in these conversations, and what people ask AI to do with them.
"Is our brand in the AI conversation at all?"
11 countries, one protocol. The same questions, asked the same way, in every market, so when the answer differs between Mexico City and Berlin, you know it's the people, not the method.
Bring your question
Not "Participant 3 said". Every answer arrives as a conclusion, what it means for your brand, and the receipt behind it, so the person who never reads research can act on it, and the person who does can check it.
STRONG 14 conversations · 6 countries · newest Aug 2026
STRONG 22 conversations · 9 countries · newest Aug 2026
EMERGING 9 conversations · 5 countries · newest Aug 2026
"We're deciding X. What does the research say?" That's the whole interface. Type it in plain language and get the conclusion, the receipts, and an honest "we don't have that yet" when it's true. No report to read first. No researcher required.
Tell Kinmat about your brand once (category, markets, what you're deciding this quarter) and every answer lands on your business, not on "consumers in general".
We're adding an AI assistant to our banking app in Mexico. What should it never do on day one?
Three lines it shouldn't cross on day one:
1 · Don't move money on its own. People want it to explain, compare, remind. Not act.
2 · Don't pretend to be a person. A bot that says it's a bot keeps trust. A disguise loses it.
3 · Don't ask for more than the job needs. Every extra permission reads as a data grab.
Not in the conversations yet: how people feel about voice instead of text. Added to the next round.
Marketing doesn't live in a research tool. So Kinmat works two ways: ask it a question, or hand it the work you already have. And it travels as a link and as a source your own AI can read.
We're opening the AI in Daily Life study to a first wave of marketing and insights teams. No seats, no report to wait for, no six-figure retainer. Access to the living study and everything it can answer, from day one.
Not sure yet? Book a 20-minute walkthrough and bring a real question. You'll see it answered with receipts before you decide.
11 countries · one protocol · growing every month
Interviews by senior researchers from the UX Alliance network · limited first wave
Kinmat's conversations are conducted by senior researchers from firms in the UX Alliance network: independent research firms that have done fieldwork for the world's biggest brands since 2005. This is not a panel. It's people who live where your customers live.
For twenty years we ran field research for some of the biggest brands in the world. And we kept watching the same scene: a marketing team about to make a real decision, with real money, asking a chatbot about their customers, because the real answer was two months and $40,000 away.
We didn't build Kinmat to replace your agency, or to add another AI that sounds confident. We built it because the conversations already exist, the researchers already exist, and the only thing missing was a way to put real human answers in front of the people deciding, in the time they actually have.
If an insight can't show you who said it, you shouldn't build a campaign on it. That's the whole company.
The Kinmat founders
No. It's a living study you ask. Answers come back as a conclusion, what it means for your brand, and the receipts behind it, shareable as a link instead of a PDF nobody opens. If you want the long read, your research team can go all the way down to the transcript.
No, and that's the point. Kinmat is 100% qualitative: recorded, human-moderated conversations. Numbers tell you what people do. This tells you why, in their words. Use both; only one of them is in your competitor's dashboard.
The study is about how people live with AI: in their money, health, food, work, learning and relationships. If you're deciding anything about AI in your product, your service or your message, it's about your customers. And the brand & category view shows you where your name comes up unprompted.
Senior researchers from firms in the UX Alliance network, each in their own country and language, following one shared protocol across 11 countries. Every conversation carries its researcher, method, place and date.
No. Plain-language questions, plain-language answers. Your insights team will love the transcripts and the coding, but nobody needs to translate for the rest of you.
It stays yours. We delete what you upload once it is no longer needed. We never sell your information, and we never use it to train models.
$3,000 per quarter, flat, for your whole organisation. No per-seat fees. Introductory price for the first wave. Request access and we'll confirm the same day.
They're already in Kinmat, alongside thousands of moments from real conversations in 11 countries, and more arrive every month. A different name every time. The same truth every time.
Prefer a conversation? Talk to a founder on WhatsApp · LinkedIn
