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.
How does the facial coding and eye tracking actually work in practice — is it through a browser plugin, a mobile app, or do participants need special hardware?
Mira
@nkurumak28108 No plugin, no app, no special hardware. Everything runs in the browser; participants just grant access to the webcam and microphone when they join. There is a quick calibration check before the session starts to make sure lighting and camera positioning are good. Works on any standard laptop or desktop webcam.
The facial coding during interviews actually caught a reaction I would have completely missed reviewing the transcript alone. Seeing emotion data layered with what people said felt like a real research upgrade, not just another AI wrapper.
Mira
@memetjr2q Thank you for putting it through a real test. That gap between what the transcript says and what the face shows is exactly what Mira is built to catch, most tools only ever see the words.
How does the facial coding and eye tracking piece actually work in practice, do participants need to opt in and run any special setup on their end?
Mira
@ezelsukut Everything runs in the browser, no downloads or setup. Participants grant camera access, run a short calibration (only for eye tracking), and the session starts on a standard laptop or phone with in-built cam. Fully consent driven.
The facial coding and emotion layer actually feels different from typical survey tools — I ran a quick concept test and the sentiment data picked up nuances I usually miss in write-ups.
Mira
@gllfevp Thank you for testing it on a real concept. The hesitation read is the whole point. People rarely say "I'm unsure" out loud, but the face shows it, and that gap between what's said and what's felt is exactly what Decode is built to surface
How does the facial coding and eye tracking actually work in practice with participants who are on their phones or in different lighting conditions? Curious how reliable that data really is across such a massive global panel.
Mira
@miratngl8j5s Great question Miraç.
A few specifics on how we handle this:
Phones: Mira currently runs in-browser on desktop, participants join via a standard webcam, not a mobile camera. We found controlled desktop sessions produce more reliable facial signal than mobile, where camera angle and movement vary too much. For global panel recruitment, participants are briefed on the setup requirement before joining.
Lighting: Before every session, participants go through a quick calibration check, the system guides them to adjust their face position and lighting until confidence scores are acceptable. Frames that fall below our threshold are automatically excluded from the emotional signal rather than being interpolated. We never guess on weak data.
Global reliability: Our models are trained across demographic groups and we apply cultural normalisation benchmarks to reduce bias in emotional scoring. The signal is most reliable when used comparatively, patterns across a study group, rather than reading single participant moments in isolation.
Honest caveat: webcam-based tracking is less precise than lab infrared. For questions like where attention went first or whether someone showed hesitation at a specific moment, it works well. For pixel-level gaze accuracy, a lab setup still wins.
Happy to go deeper on any of this.
Timbal AI
When Mira spots a say/feel mismatch and digs deeper on its own, how do you keep that follow-up from leading the participant? A moderator reacting to visible confusion can easily plant the doubt rather than uncover it. Is the probe neutral by design, or tuned per study?
Mira
@david_vilalta David, this is the question we lost sleep over. A leading probe is almost worse than no probe — it plants the narrative.
A few design decisions we made:
Probes are behaviorally triggered but linguistically neutral. When Mira detects a signal — a hesitation, an emotional shift, a response that doesn't match the facial/voice pattern — it doesn't say "you seemed confused." It says something like "You paused there — tell me more about what was going through your mind." The trigger is emotional, the language is open.
No interpretive language in the probe. Mira never names the emotion it detected back to the participant. It asks outward, not inward. This is a deliberate constraint built into the probing engine.
Tunable per study type, not per participant. Concept testing probes are calibrated differently from usability or brand research — because the type of signal differs. But within a study, every participant gets the same probe structure. That keeps cross-participant comparisons valid.
Probe depth is capped. Max 2 levels of follow-up on any one signal — so the interview doesn't become an interrogation.
The honest caveat: no probe is perfectly neutral. But compared to human moderators who vary tone and word choice across 40 interviews — Mira is at least consistently neutral. Happy to show you a live session → https://www.entropik.io/book-demo
Timbal AI
What happens when it works fine and every brand in a category runs creators against the same queries, does it become an arms race where the UGC cancels out, or is there a ceiling on how much citation share you can actually buy back?
Mira
@david_vilalta Sharp concern, and one worth taking seriously.
The short answer: the questions might converge, but the insight won't — and here's why.
The emotional layer is proprietary to each brand's product. Two competing brands can ask participants the same question about their respective checkout flows. The language of the answers might look similar. But Mira's facial coding and voice emotion data will surface where frustration spikes, which exact moment trust drops, what triggers genuine delight — and that's product-specific. You can't benchmark emotion.
Research memory compounds differently for each brand. Mira builds cross-study intelligence over time — recurring themes, longitudinal shifts, segment-level emotional patterns. A brand that's been running studies for 12 months has a fundamentally different starting point than one that just launched. That proprietary research memory is not replicable even with identical questions.
Speed becomes the moat, not secrecy. If every brand could run the same study, the winner is the one that runs it first, iterates fastest, and acts before the category moves. Mira compresses weeks of manual research into hours. That time advantage is where competitive edge lives.
The ceiling isn't on insight quality — it's on how fast you can act on it. That's what we're really building for → https://www.entropik.io/book-demo