FAQ

PersonaChoice is a marketing research tool that combines AI with theories from behavioral economics and psychology.

It analyzes both the “unstated preferences and intuitive reactions” and the “stated reasoning (logic)” of consumers — things that were hard to capture with traditional surveys — so you can simulate the effectiveness of product ideas and ad copy in advance.

The biggest differences are “eliminating bias” and “speed.”

Surveys of real people are affected by noise such as answering just for the reward or wanting to “look good.” PersonaChoice’s AI personas do not carry these biases and act frankly based on their configured personality and values. And a survey that normally takes several days can be completed in minutes with PersonaChoice.

Yes, creating an account is free.

You receive 3,000 points at sign-up, so you can try small tasks for free right away. For ongoing use, please consider our cost-effective monthly plans or point purchases.

PersonaChoice is a marketing research tool that simulates the “tendencies” and “preference rankings” of human decision-making.

In comparisons against published real-world survey data, a high degree of agreement in ranking and tendency has been confirmed. For the specific validation content and results, please see the Case Studies page.

It’s the difference between asking “one smart consultant” for an opinion and having “1,000 diverse consumers” vote.

A general AI like ChatGPT returns a single “logical answer” that averages the vast information on the internet. Like a capable consultant, it’s good for brainstorming, but it can’t tell you the “market’s reaction.”

PersonaChoice, on the other hand, simulates 100,000 consumers with different personalities and values. The decisive difference is that it can visualize the human-like “hesitation” and “split votes (%)” that vary person to person — like “not logical, but I somehow like it” or “the quality is good, but it’s too expensive to buy.”

In this service, we classify the personas who act as respondents as follows.

• ●
  Persona (general term): the collective term for the AI characters that take part in a simulation.
• ●
  AI persona: a standard AI persona prepared in advance by PersonaChoice, with general attributes (age, gender, occupation, etc.). You narrow them down with conditions such as “women in their 20s, Tokyo, service industry.”
• ●
  Synthetic audience: a “survey population” composed of many AI personas. There is a standard synthetic audience that is ready to use immediately based on official statistics (such as the census) even without your own data, and a company-specific synthetic audience built to reflect a company’s membership base for companies that have member/customer data (enterprise; please contact us).

It is the unit of a survey — such as a “questionnaire” or an “A/B test” — that you ask the personas to complete.

Within one task, you set the questions and options, select the target personas, and request their responses. You then obtain analysis results from those responses.

You can use two modes depending on your goal.

You can set the number according to your plan and point balance.

To hear a small number of qualitative opinions, around 20–50 people is fine; for statistical data, we recommend a scale of 100–1,000 people. The Ultra plan can run tasks with up to 5,000 respondents.

Yes. In addition to question text, you can upload images and show them to the personas. It’s ideal when you want to analyze intuitive reactions to visual elements, such as “A/B testing package designs,” “comparing logos,” or “selecting ad banners.” You can use it in both Quick mode and Advanced mode.

After a task completes, a detailed report is generated including the following data. (Report content is currently provided in Japanese.)

• ●
  Quantitative data: the number of votes per option, visualized with charts and tables by respondent age group, gender, and so on.
• ●
  Qualitative data: free-text reasons for “why they chose that option.” Each persona gives concrete reasons based on its own personality and values.
• ●
  Psychological analysis: an analysis of whether “logic (rational judgment)” or “intuition (emotional judgment)” was dominant.

Yes.

From the “Download PDF” button on the task report screen, you can download a PDF that compiles the task report. You can use it as-is for internal meeting materials or client proposals.

• ●
  Please check that your email address and password are correct.
• ●
  Please check your spam folder.
• ●
  If it still isn’t resolved, please contact us via the contact form.

Analysis usually completes in a few minutes, but it may take longer depending on server load and the number of personas. If the status does not change for a long time, please contact support.