Start with nothing
Candor searches for published data about the population you described: government statistics, industry reports, peer-reviewed research, professional association data. Every signal it finds is tagged with where it came from.
Describe who you want to understand. Candor builds synthetic participants from real evidence, runs research with structured methodology, and delivers findings you can act on. Three steps. An hour, not weeks.
Candor is secondary research that makes primary research sharper. It’s designed to fit into the work you’re already doing, not replace it.
Step 1
Tell Candor about your target audience: who they are, what they do, what industry they’re in, whether they’re business buyers or consumers.
Candor searches for published data about the population you described: government statistics, industry reports, peer-reviewed research, professional association data. Every signal it finds is tagged with where it came from.
Customer interviews, support tickets, survey data, market reports, internal briefs. Candor prioritizes your documents over web research, so participants reflect your specific customers and market, not just the broader population.
Candor reads through every source and pulls out specific claims about your audience: how they behave, what they believe, what they care about. Each claim is labeled with where it came from, so you can see how much to trust it.
Candor groups your audience based on what actually drives their decisions, not just demographics. The grouping adapts to your context:
Grouped by role, decision-making authority, openness to adopting new technology, and other factors that shape how organizations evaluate and buy.
Grouped by personality, beliefs, motivations, and life situation that shape how individuals decide and spend.
Within each group, Candor creates individual participants. Every participant gets:
Participants within the same group are deliberately diverse. Instead of generating the average person in a group, Candor samples personality traits and biases from distributions, so two participants in the same group can hold different views and react to the same idea differently. You review participants before running any research.
Step 2
Describe what you want to learn and Candor recommends a study type based on your question:
Open-ended interviews that go deep on behavior with no stimulus introduced.
Test whether one or more problems are real, frequent, and painful enough to act on. The hypothesis is withheld until after an unprompted check, so participants aren't led toward your answer.
Introduces the idea early and probes for comprehension, objections, and perceived value. Includes sub-types for specific needs: an idea or feature test, a name test, a value prop test, or a messaging test.
Confirms value before price is introduced, then explores willingness-to-pay and price sensitivity.
The question design, the depth of follow-up, and the sequencing all change depending on the study type. Discovery interviews stay wide open. Concept tests introduce the stimulus early and probe for whether people actually understood it before asking what they think. Validation interviews withhold the hypothesis to avoid leading the participant. Price tests establish value before introducing numbers. This isn’t one generic interview template with different labels.
Candor conducts the interviews automatically. Its interviewing methodology has been infused with research rigor from decades of research conducted across academia, UX, consulting, and venture building. The interviewer adapts to what each participant says, follows up on the threads that matter, and knows when a topic has been explored enough to move on.
Before every interview, the system loads each participant's full profile: their personality, background, beliefs, and behavioral patterns. If anything fails to load, the interview doesn't start.
During the interview, participants are continuously checked against their profile so they stay in character and don't drift toward a generic, agreeable version of themselves.
The interviewer probes for objections regardless of how positive the conversation has been, and checks whether the participant actually understood the idea before asking what they think.
Every question is screened before it's asked. Leading questions, loaded questions, yes/no questions, and questions that reveal the solution too early are all blocked.
A separate system watches for participants drifting out of character, making up facts, agreeing too easily, or contradicting themselves. Problems are corrected silently and flagged in the synthesis.
Step 3
After the interviews, Candor reads through every transcript and pulls out what matters: what people do, what frustrates them, what they want, what they believe, and how they make decisions. It finds patterns across participants, not just individual responses.
Candor doesn’t limit itself to what you asked about. If something unexpected comes up across multiple interviews, it surfaces it. New problems you didn’t think to test for, unanticipated objections, needs that weren’t in the interview guide. These often turn out to be the most valuable findings.
Findings are grouped by theme and ranked by how many participants raised them, broken out by group.
A clear answer on each problem you tested: validated, partially validated, not supported, not applicable to this audience, or inconclusive. Each comes with the evidence behind it.
What worked and what didn't, broken out by group. Each idea gets a clear verdict, from pursue to drop, with the reasoning behind it. If people didn't understand the idea, that's flagged as a bigger issue than whether they liked it. For name, value prop, and messaging tests, you also get a direct comparison across the options you tested.
How people perceive value, where willingness to pay breaks down, and which pricing structures feel right versus wrong. Candor surfaces both specific price reactions and broader patterns across participants.
Tensions where participants genuinely disagree are surfaced and broken out by group. You see which groups split and why.
Unanimous agreement is treated as a caution, not a conclusion. Consensus from synthetic participants can reflect model behavior rather than a real pattern, so suspiciously clean agreement is a signal to probe further with real participants.
When you're entering a new market or audience and don't know enough to write a good interview guide, Candor gives you a starting point. Run a discovery study with synthetic participants to surface the themes worth exploring, then go deeper with real people.
You've talked to a handful of real people and learned something. Now you want to pressure-test it, go deeper on a specific thread, or see how a concept lands with a broader audience. Upload your transcripts and Candor builds participants grounded in what you've already learned.
Niche B2B audiences, hard-to-reach populations, tight timelines. When you can't get the right people on a call fast enough, Candor gives you signal to work with while you recruit.
Run the full sequence (discovery, validation, concept testing, pricing) with the same synthetic audience in hours instead of months. Then validate the decisions that matter most with real participants.
Most studies run in 4 to 8 hours of elapsed time, with roughly an hour of your active attention spread across that window. Your audience and its synthetic participants generate together in one background run of 25 to 35 minutes (it's queued, you can close the tab). Interview guides generate in a couple of minutes. Live interviews happen in real time. Auto-interviews run in the background and finish based on participant count and interview depth. Synthesis runs automatically a few minutes after the last interview completes. The bottleneck is your time spent reading and reviewing, not Candor's processing time. End to end, you go from a research question to a structured report inside a single workday, instead of the weeks a panel study would take.
Your uploaded documents come first: customer interviews, survey data, market reports, or any research you already have. Candor supplements these with web evidence retrieval, searching for corroborating data about your target market. Where a signal has a source, it shows whether that source was your own document or a public web page, and each participant attribute is labelled grounded, derived or assumed.
No. Candor is built for UX research and insights leaders who want to scale their research capacity, plus product teams, innovation teams, consultants, and founders who need customer signal without a dedicated research function. You don't need research training. Describe your audience in plain language, choose a study type, and Candor handles the methodology. Auto-interview manages question flow, coverage tracking, and signal extraction on its own. The synthesis report is structured to drive product decisions, not academic analysis.
Yes. Synthetic participants are persistent within a study. When you return to interview one again, they remember everything from prior sessions: specific stories they told, decisions they described, and how their views evolved. Show them a new concept and they'll react with full context of your conversation history.
Four study types, and concept testing comes in four kinds. Problem discovery explores problems you don't know about yet. Problem validation tests specific problems against what participants say and do, with a verdict on each. Concept testing evaluates what resonates and what needs rethinking. Price testing examines value perception, anchoring, and willingness to pay. Inside concept testing, a standard test compares whole ideas, while a name test, a value proposition test and a messaging test compare options for one concept you already have. The synthesis report adapts to whichever type you run. See the full set on the use cases page.
Start with a free project, or walk through Candor with us first.