Skip to content
Kapari
How it works
Who it's for
Executives HR Agencies Consultants
Examples Proof Security Pricing Demo Version française Log in Start your 30-day free trial
Understanding the market

Customer digital twin: what it is and what it is worth

A customer digital twin is a copy of a customer, created by an AI from data about them, that you question in their place. Useful for testing a message. To know how an announcement will land, there is a better-suited tool.

The promise

Question your customers as often as you like

The word comes from industry, where engineers test a repair on the digital copy of a turbine before touching the real one. Applied to customers, it promises the same thing: test ten messages instead of one, the same day, with no study to organize.

Before you buy

The question that matters: where does the copy come from?

The best known result comes from Stanford (2024): after a two-hour interview with a person, their copy matches 85% of their answers to an opinion questionnaire. A copy built from a plain customer database interviewed nobody, and that number does not apply to it. So ask the vendor what data their twin rests on, and whether their score was measured on your customers.

Twin or decision test

Your question is not "what would this customer say?"

When you raise prices, close a site or change an organization, the real question is: how will my customers, my employees and my partners take it, together? A twin answers for one customer. Kapari brings varied voices together around your decision and shows you who goes along, who pushes back, and what to fix before you announce.

On 100 real past decisions, Kapari finds on average 77% of the criticism raised at the time. See the exam · The digital twin applied to a decision

For the curious: the sources
  • Park et al., Generative Agent Simulations of 1,000 People, Stanford, 2024: 1,000 people interviewed for two hours each, 85% fidelity; copies built from simple demographic descriptions are more stereotyped.
  • Bisbee et al., Political Analysis, 2024: from one person to a group, a correct average can hide wrong answers.
  • The wider picture: our guide to synthetic audiences.

Try it on your next announcement.

Write down what you plan to announce. In a few minutes, you see who goes along, who pushes back, and what to fix.