Skip to content
Dori

BUILD WITH YOUR CUSTOMERS IN MIND

Synthetic customer testing for builders

Synthetic customer testing uses AI participants to explore a product from different customer perspectives. Dori helps builders find confusion, friction and trust issues before sharing a product with real users.

THE SHORT ANSWER

What is synthetic customer testing?

It is a way to rehearse a product experience with simulated people. You give AI participants a customer context and a task, then inspect what they do, where they struggle and what the evidence suggests changing. The findings help you form hypotheses for your next iteration.

When to use it

Use it when you have a landing page, prototype or working web product and want an early view of how different people might experience it. It can help solo builders, product teams and builder agents examine clarity, navigation and trust before a human research session.

How to approach the test

  1. Set a decision and a task

    Choose a specific question, such as whether a new visitor understands the product before signing up. Give participants a realistic goal rather than asking them to find every possible problem.

  2. Choose relevant perspectives

    Create participants with the needs, context and familiarity you want to explore. A time-constrained first-time visitor can reveal different friction from an experienced evaluator.

  3. Review behaviour and evidence

    Inspect the journey, findings and available screenshots or recordings. Check that a finding is supported by a visible action or response rather than an assumption about a whole customer segment.

  4. Change one thing and test again

    Turn a finding into a concrete product change. Revisit the task, then use real customer research and product data to validate the decision.

ILLUSTRATIVE EXAMPLE · NOT A LIVE TEST RESULT

From an observation to a product change

task
Understand the product and decide whether to create an account.
participant
A first-time builder evaluating a new testing tool.
observation
The participant opens the pricing section, returns to the hero and looks for an example report before choosing Sign up.
evidence
Illustrative sequence: pricing → hero → search for a report example. A real finding should link to the relevant steps and captured evidence.
interpretation
The visitor may need to see what a test produces before committing to setup.
change
Show a clearly labelled sample report beside the primary call to action.

What to look for

  • Value proposition and next-step clarity
  • Terminology and navigation
  • Trust signals and expectations
  • Differences between participant perspectives

What this test can and cannot tell you

AI participants do not represent a statistically sampled customer population. Their actions depend on the task, persona context and model. A repeated finding is a reason to investigate, not a measured prevalence rate, conversion forecast or proof of what real customers will do.

Common questions

Does Dori replace real customer research?

No. Synthetic tests can surface questions earlier. Interviews, usability sessions with people and behavioural data are still needed to confirm customer needs and product outcomes.

What should I test first?

Start with one decision and one short task: understand the product, find a price, sign up or reach a first useful result. Narrow tasks make findings easier to act on.

How should I judge a finding?

Look for a clear observation, the step where it happened and supporting evidence. Separate what the participant did from the interpretation and the change you want to try.

Build a better next version.

Explore Dori and create your account. Test execution depends on current availability and account access; check your account before planning a study.

Create your account