Unlike AI tools that stop at interview + transcript, Mira is a full AI researcher — plans studies, recruits globally (100M+ panel, 120 countries), runs dynamic interviews with intelligent probing, and uniquely captures what participants say AND feel via real-time facial coding, voice emotion AI, and webcam eye tracking. Extracts themes, generates insights, and produces research reports automatically. 17 patents. 70+ languages. Trusted by Unilever, Nestlé and 150+ global brands. $25M Series B.
Forum Threads
The say/feel gap in research; how big is the problem actually, and what do you do about it?
There's a well-documented phenomenon in consumer and user research that most practitioners know intuitively but rarely name directly: people don't report their experience accurately.
Not because they're dishonest. Because self-reporting is hard. In the moment of an interview, participants are performing a version of themselves. They round off hesitation. They describe their behaviour more charitably than it actually was. They say "yes, I'd probably use this" when what they felt was closer to "maybe, under the right circumstances, if the price were different." The social pressure of being in a conversation, even with an AI, shapes what gets said.
This is what we call the Say-Do Gap. The distance between what someone tells you and what they actually feel or do.
It shows up everywhere. In concept tests where participants say they love a product they'd never buy. In usability sessions where someone says "this makes sense" while visibly struggling. In brand perception studies where stated attitudes don't match purchasing behaviour.
What do transcripts miss that sends you back to the recording?
Nine years of working closely with insights teams across CPG, BFSI, and tech has given us a consistent observation: the transcript is where researchers start, but the recording is where they go when the transcript isn't enough.
They go back to hear how someone said something. To check whether the pause before an answer was genuine uncertainty or just thinking time. To see whether the person's expression changed when they described a feature they claimed to like. To catch the moment when engagement visibly dropped when they stopped leaning forward, when their eyes moved away from the screen even while their words stayed polite and positive.
That gap between the transcript and the recording is where a significant amount of research insight lives. And it's also where the most time gets consumed in manual analysis. Researchers who run 20 or 30 interviews know what it's like to sit through hours of footage looking for the three or four moments that actually matter.
This observation that the transcript captures what was said but not how or with what underlying feeling is fundamentally what drove us to build the emotion and behavioural layer into Mira. Not to replace researcher interpretation, but to give researchers a way to find the right moments faster.
I'm genuinely curious: what's the signal you find yourself wishing you had after you've finished going through a set of interviews? What consistently drives you back to the recording that the transcript alone doesn't give you? And have you found any workarounds annotation practices, tagging systems, collaborative review setups that help you capture more of what the transcript loses?
What would actually make you trust an AI moderator's findings?
We launched Mira on Product Hunt today, and the conversations in the comments have been the most interesting part of the whole day.
Researchers asked about probe neutrality whether the follow-up questions an AI asks mid-interview can lead a participant, rather than uncover them. They asked about cultural calibration, whether emotion models trained largely on Western data can accurately read a participant in Jakarta or Nairobi. They asked whether the say/feel mismatch gets surfaced as raw evidence or quietly resolved into a single confidence score.
These are serious questions. And they made me realize something: the bar for trust in AI-moderated research is fundamentally different from other AI tools.
If a writing assistant gets something wrong, you catch it before you publish. If an AI coding tool hallucinates, your tests fail. But if a research tool misreads how participants felt during a concept test, and that feeds into a product decision, the error is invisible. By the time the product ships and the market responds, the research moment is long gone.
The facial coding and emotion AI combo is wild — usually I get a wall of quotes to sift through, but here it flagged where respondents actually felt something. Made the insight pull way faster than expected.
Mira
@doanj7hh "Flagged where respondents actually felt something" is the signal we built toward — not a wall of quotes, but the specific moments that matter emotionally. Glad the insight pull was faster than expected.
Mira
Hi all,
Marketing lead at Decode here.
I have spent the last few months working closely with this product to build the launch. The thing that struck me most: most AI interview tools stop at the transcript. Mira treats that as the starting point.
A few capabilities people miss: AI follow-up probing that automatically asks "why?" and "tell me more" mid-interview based on what the participant actually said. AI highlight reels that automatically pull emotional moments, so you do not have to share 40-minute recordings with stakeholders. Cross-study intelligence that finds recurring themes across multiple research projects over time.
The multimodal layer — emotion AI on top of the interview — makes findings more defensible. You are not just quoting a participant. You are showing what they felt when they said it.
Happy to answer questions about the product or how we built the launch.
How does the facial coding and emotion AI handle privacy and consent across different regions, especially with GDPR and other strict regulations?
Mira
@ufukaktugba
Hi Ufuk, great question.
Consent is handled per session: participants see a clear explanation of what data will be captured (facial, voice, eye) before any session begins and must actively opt in.
On GDPR: we are fully GDPR compliant. Data is processed within compliant infrastructure, participants can request deletion, and researchers control retention policies. We also have SOC 2 Type II certification.
For APAC markets we follow regional equivalents (PDPA in Singapore, PIPL in China). If you are evaluating for a specific region, happy to go deeper.
BetterClaw
This is a really interesting take on AI-powered research.👏🏻
Going beyond transcripts to capture emotions and behavioral signals could unlock much richer insights for product teams. Curious, how do you balance those advanced features with participant privacy and consent during interviews?
Mira
@worksforme Hi Laiba — the balance is built into the session flow itself. Every participant gets a clear explanation of what will be captured before they join — webcam access for facial and eye data, microphone for voice emotion. They can decline any of these and still participate, with the system adapting to use only the available signals.
No facial recognition is used anywhere — we only measure expressions, not identity. Data is never sold or used for external training without explicit permission. Researchers control retention and deletion.
Happy to walk through the full consent flow if that would help.
How does the facial coding and emotion AI actually perform across different cultural contexts, and do participants need any special setup or permissions on their end for it to work smoothly?
Mira
@duyguridapagq Our emotion AI is trained to understand the demographic difference while predicting emotions using the face. Later we do a few normalisations based on the benchmarks we have collected to give you unbiased scores and emotion data points. We just need permission to use the microphone and the webcam.
How does the facial coding and emotion AI actually perform across different cultures since expressions and emotional cues vary so widely globally?
Mira
@zgreg0z Our models are trained across demographic groups rather than on a single population, and we apply normalisation based on cultural benchmarks we have collected to reduce bias in emotional scoring. The goal is to measure genuine reactions, not to apply one cultural standard globally. We also let researchers set study-specific calibration. Emotion expression does vary across cultures and we do not claim perfect universality — it is an ongoing area of work. For global studies we recommend always combining the emotional signal with the transcript and not reading it in isolation.
Mira
Product Manger at Decode here
One thing that always fascinated me about qualitative research: researchers don't just analyze what people say. They spend hours replaying interviews to understand how they said it.
A micro-expression of disgust at the pricing slide, a long pause before "I'd probably use it," hands that stopped moving the moment they said they were "comfortable."
That's the invisible layer of qual research. The part that turns a quote into an actual insight.
Mira is built around that gap, it conducts interviews in 75+ languages, probes emotionally in real time, and simultaneously captures vocal hesitation, facial emotion, and where attention went. So researchers stop losing those moments to memory and manual replay.
The goal isn't to replace researchers.
It's to give them superpowers, so they can spend more time discovering why people behave the way they do, instead of manually reviewing hours of recordings.
Happy to answer any questions about how any of this works.