Aaru and Candor both use AI to stand in for customers, but they answer different questions. Aaru simulates thousands of people answering a survey and tells you how the answers split. Candor interviews synthetic participants (often called synthetic users) one at a time and tells you how a type of customer reasons, and why.
| Dimension | Aaru | Candor |
|---|---|---|
| What you ask | Survey-style questions: single choice, multiple select, ranking and free response, with stimuli, hypotheticals and branching | Open questions in one-to-one interviews, run by you or by Candor's automated interviewer, which asks follow-up questions |
| What you get back | How answers split across a population, with cross-tabs by audience and exports to spreadsheet, slide deck and PDF | Interview transcripts and a written report on how participants reason, where they agree and where they split |
| Who answers | A population of AI agents that matches the audience in aggregate. Aaru's example runs use 10,000 agents per audience. | Synthetic participants, each built from evidence about a type of customer. Up to 24 per audience. |
| What it's built from | Public records such as censuses, licensed data such as anonymized transactions, store visits and search demand, plus your own data if you have it | Your uploaded research first, then web research. Personality traits come from published population data for the participant's region and occupation. |
| Published accuracy | A study of 2,993 survey questions with an average gap of 3.53 points per answer option, and a blinded EY test with a median correlation of 0.90 | None yet. Candor hasn't published a test of its output against real human answers. |
| Pricing | Not published. You contact Aaru's team. | Free: $0 a month, $10 per automated interview. Standard: $75 a month, $7 per automated interview. Interviews you run yourself are free. |
Aaru facts come from aaru.com, its simulation page, EY case study and validation study, checked on September 26, 2026. If a row is out of date when you read this, tell us and we’ll fix it.
The difference is the shape of the answer. Aaru gives you a number for a population: what share of value shoppers would try a store brand at $2.99, say. Candor gives you a conversation with a participant: why they hesitate, what they compare it with, and what would change their mind.
Aaru builds a population of AI agents whose traits, such as age, income and location, match the audience you describe. Each agent answers your questions, and Aaru adds up the answers. Its simulation page says a result is a spread of likely behavior across the population, not a prediction about any one person.
Candor builds a small audience, up to 24 participants, from evidence about a type of customer. You interview each one, or let Candor’s automated interviewer do it, and follow the reasoning wherever it goes. Neither approach is better in general. A survey answers “how many?” and an interview answers “why?”
Published accuracy tests. Aaru’s September 2026 validation study compared its simulations with published results for 2,993 questions from 186 studies across nine industries. The average gap was 3.53 percentage points per answer option, or 2.78 after allowing for sampling noise in the surveys. It also compared brand rankings with real card-purchase data and ranked 92.4% of brand pairs in the right order, against 77.8% for published survey answers.
A blinded test with EY. In its EY case study, Aaru recreated EY’s 2025 Global Wealth Study, a survey of 3,600 affluent investors in more than 30 markets, without seeing the results. The run took a day. Across 53 questions the simulation matched the survey with a median Spearman correlation of 0.90, a measure of how closely two sets of answers agree in order, where 1.0 is perfect. Where they differed, Aaru argues the simulation was closer to what investors actually do than to what they said.
Scale and reruns. A population can hold tens of thousands of agents, and the same population can be rerun after a rate change or a competitor’s move. That suits market sizing, segmentation and scenario planning, where the question is how a large group shifts.
Enterprise backing. Aaru’s about page quotes leaders at EY, Accenture Song and Interpublic Group, and lists offices in New York, Singapore and San Francisco. Candor is newer. If published accuracy figures and an enterprise client list are part of your case, Aaru has both and we don’t yet.
Candor gathers evidence about your audience before anyone exists to interview: your own research first, then web research. It groups the audience from that evidence and builds each participant from it. Personality traits are drawn from published population data for the participant’s region and occupation.
The details about who a participant is, how they behave, what they believe and how they talk each carry a label. Personality and bias scores don’t. Grounded means there’s something you can check behind it. Derived means it follows from the evidence without a source stating it outright. Assumed means Candor filled it in and says so. Then you interview. Participants remember earlier conversations within a study, so you can come back and dig further. How it works walks through the full process.
Aaru has published its accuracy against real survey results. Candor hasn’t yet published a test of its output against real human answers, so we make no accuracy claim for it. Candor’s 24 participants also can’t tell you what share of a market will do something. Treat its findings as a map of what to test with real customers next, not a forecast.
What Candor gives you in return is depth. A survey question has to be written before the answers come in, so it can only measure the options you thought of. An interview can follow a surprising answer and find the reason you didn’t know to ask about. Early in a decision, that reason is often worth more than the percentage.
Price is the other difference. Aaru doesn’t publish prices, and you start by contacting its team. Candor has a free plan and charges per automated interview, with no contract, so a small team can run a study this week. The pricing page has the details.
Choose by the question you need answered. Many teams need both kinds of answer at different points in the same decision.
This comparison is one of several. See Candor vs Brox for a tool that builds digital twins of real people, Candor vs Synthetic Users for another interview-style platform, Traditional vs synthetic user research for survey research with real people, or the full comparison hub. The synthetic respondent entry explains survey-style and interview-style tools in more depth, and what is synthetic user research covers the whole category.
Aaru is a company that simulates how groups of people will behave. You describe an audience and write survey-style questions, and Aaru builds a population of AI agents that matches that audience and answers them. Its example runs use 10,000 agents per audience. Aaru builds each population from public records such as censuses, licensed data such as anonymized transactions and store visits, and your own data if you have it. It sells to large organizations: its site quotes leaders at EY, Accenture Song and Interpublic Group.
For some jobs. Both use AI to stand in for customers before you spend on real research, so teams weigh one against the other. They answer different questions, though. Aaru predicts how a whole population's answers will split across the options in a survey. Candor runs one-to-one interviews with synthetic participants to learn how a type of customer reasons and why. If you need numbers for a large population, Aaru fits. If you need to understand a problem, a concept or a price in depth before real interviews, Candor fits.
Aaru doesn't publish prices. Its site has no pricing page, and you contact its team to start. Candor publishes its pricing. The Free plan is $0 a month with $10 automated interviews. Standard is $75 a month with $7 automated interviews. Interviews you run yourself are free on both plans. A 24-participant study on Standard costs $243 in its first month, monthly fee included.
Aaru has published more accuracy evidence than most tools in this category. In a September 2026 study it compared its simulations with published results for 2,993 survey questions. The average gap was 3.53 percentage points per answer option, or 2.78 after adjusting for the sampling noise in the surveys themselves. Separately, EY asked Aaru to recreate its 2025 Global Wealth Study without seeing the results. Across 53 questions, the simulation matched the survey with a median Spearman correlation of 0.90. That measures how closely two sets of answers agree in order, where 1.0 is a perfect match. It isn't the same as being right 90% of the time.
No. Candor runs one-to-one interviews, either by you in a chat or by Candor's automated interviewer, which asks follow-up questions. An audience has up to 24 participants, so Candor's output is transcripts and a written report on how people reason, not percentages for a population. If you need to know what share of a market will pick option A over option B, a survey-style tool like Aaru, or a real survey, is the better fit.
Yes, in different ways. Aaru says customer data such as segments, purchase histories and prior research narrows its population to your decision, and that it can build one without it. Candor treats your uploaded research, such as interview transcripts, survey exports and CRM notes, as its strongest evidence. It pulls findings about a type of customer out of those files and builds participants from the findings, not from any one person in them.
Start with a free project, or walk through Candor with us first.