The Customer Interview That Almost Never Happened

The customer was available. The product team had questions. Then a release date arrived, the interview slipped, and somebody made the decision anyway. Ned Dwyer and PJ Murray knew this sequence because they had lived it. In their 2021 launch note for Great Question, Dwyer admitted that research at their previous startup was improvised and sometimes skipped to get a release out the door. Their new company began with an unusually plain proposition: make the useful conversation easier to arrange than to avoid.

The phrase “talk to customers” enjoys the status of common sense in software. It is also remarkably easy to ignore. A team has to find the right people, obtain consent, screen them, find time on several calendars, pay an incentive, record what happened, and make the result legible to colleagues who were not in the room. Each step is small. Together they are a convincing excuse to rely on instinct.

Where the interview goes to die

Great Question treats that chain as one job. A company can build an opt-in panel of its own customers, segment people by relevant traits, invite a suitable group, schedule sessions, run interviews or surveys, and store recordings and findings in the same place. It also supports prototype tests, focus groups, card sorting and tree testing. An external panel can fill gaps when the company's own customers are the wrong audience or too scarce.

That sounds prosaic. It should. A research program rarely collapses because nobody invented a new methodology. More often, the list of participants lives in one spreadsheet, invitations in an email tool, calls in a video app, and insights in a deck that disappears into a drive. Great Question's competitive argument is that the joins between those tools matter as much as the tools themselves. Dovetail is a familiar repository alternative, Maze focuses on testing, and UserTesting offers a broad testing and participant network. Great Question sells the connected journey from candidate to evidence.

The idea has a quiet consequence. A participant is no longer just a name attached to one study. The panel remembers consent and prior contact; the repository remembers what people said. A product manager can ask a new question without starting the administrative work from zero. A researcher can see what has already been learned, and can set the rules for who gets access to sensitive material.

The striking number is a headcount

At Brex, the obstacle was not a shortage of curiosity. People outside the research function wanted to talk to customers, but recruiting required help from data teams, cold email and several disconnected tools. According to Brex's account published by Great Question, roughly 10 people had been running studies. In the first year after the platform came in, more than 100 did. The company also moved its research repository into Great Question, so arranging a conversation and finding an earlier one happened in the same system.

Asana tells a related story with a different clock. Its product teams had been taking about two weeks to get in front of customers. It reports that the wait fell to two or three days, while one research operations specialist supported 45 teams. Jon Dobrowolski, the product leader in the case study, had been skeptical of research tools. What changed his mind was less a particular feature than the operational effect he had seen at other companies: one person could enable many more conversations.

“Seeing what Great Question did to extend that one person to be able to seem like they were five people ... made the biggest difference.”Jon Dobrowolski / Asana product leader

These are customer-reported results, and they describe particular organizations, not a guaranteed return for every buyer. They do, however, expose the mechanism. Remove the wait between question and participant, and more people ask. Keep the answer where colleagues can find it, and fewer people repeat the same study. That is what Great Question means when it talks about democratizing research: widening access to the work while research specialists still set standards and guardrails.

What the software costs, and what it asks of you

Great Question sells software by paid seat and module. Its public self-serve plan is $129 per seat monthly, or $1,290 per seat annually, for teams of up to five paid users. Observer seats are free. Enterprise plans have custom annual pricing and add controls such as single sign-on, finer permissions and audit logs. External recruitment is an additional, usage-based cost. A 14-day free trial lets a team test the workflow before signing up.

The cost question is therefore not simply whether an interview can be scheduled with a calendar and recorded on Zoom. Of course it can. The comparison is the labor of keeping a reliable panel, managing consent, paying participants, preserving evidence and getting the next colleague to use it. A tiny team running two studies a year may find its existing tools sufficient. An organization with many product teams and frequent decisions has more chances to recover the cost in saved coordination.

There is a version of this playbook that needs no new software. Start with a small opt-in list of customers, record what each person agreed to, and tag the list by the questions your team actually asks. Give colleagues one way to request a conversation. Keep the resulting notes beside the participant record, with a link back to the recording. Measure the days from question to scheduled call. If that number stays high, the bottleneck is visible; if it falls, the habit has a chance to stick.

The approach also has prerequisites. A customer panel needs real customers willing to take part, and somebody must keep it current. A quick interview can still be badly framed; a clean transcript can still support an unwarranted conclusion. Great Question can make research easier to conduct and share, but the team still has to choose the right question and judge the evidence. The company seems aware of this tension: its pitch couples access for non-specialists with governance for researchers.

The newest colleague is a machine

In 2025, Great Question announced a $13 million Series A led by Inovia Capital, with Y Combinator, January Capital and Character Capital participating. That same year, it integrated User Interviews' external panel of more than six million verified participants. The partnership filled a practical hole: even an excellent system for recruiting your own customers cannot summon a customer type you do not yet have.

The company has since added AI moderated interviews, automatic study synthesis and a connection that lets AI assistants work with its research system. In August 2026, it described more than 112 actions available through that connection, from drafting a screener to reading transcripts. The useful test for those features is the same as for the older scheduling tools: do they make an answer easier to inspect, and does the evidence survive contact with the next decision? A polished summary without a traceable participant quote merely makes a weak conclusion travel faster.

Great Question says its customers published 14,166 studies in 2025, up 73 percent from the year before. It is an adoption measure, not proof that every study was good. Yet it explains why this company occupies an interesting corner of the software market. It is selling the repetition of a sound habit. The name is a little grand; the work is deliberately mundane. A customer has something to say. The trick is to hear it while there is still time to change the product.