A specific user attempting one clearly defined task
Ground the study in customer evidence relevant to this decision.
Sunzu for ai engineers and app builders
Your agent builds the UI.
Who checks the experience?
Move from a working interface to a better-understood user experience. Use Sunzu through a compatible MCP client to explore likely friction before the next release.
Used by
- Insights analysts
- Product managers
- UX designers
- Marketing strategists
- Copywriters
- Go-to-market leads
- Growth managers
Trusted by
Backed by
- Microsoft for Startups
- NatWest Accelerator
The work in front of you
Put user questions next to the code.
Connect your account using the Sunzu MCP setup guide. Start with one audience, one task and one research question.
- UX testing
Review one critical experience
Use UX testing to explore likely friction in a focused flow, such as first-time setup or inviting a teammate.
Explore ux testing - Concept testing
Challenge the idea before generating more code
Compare concepts with your audience and understand which direction deserves another iteration.
Explore concept testing - Depth interviews
Research from your assistant
Connect Sunzu through MCP for depth interviews, concept tests and UX tests without switching to a separate research workspace.
Explore depth interviews
Your next study
Will the intended user understand the next step?
Bring a real question. We’ll scope the audience, evidence and study together.
Start with what you have
Experience to review
User task and context
Research audience
From research to a decision
A focused research brief for your assistant
Keep the question, the evidence and the next action connected.
Ask
Review the first-time setup experience for a small-team administrator. Explore whether the invitation step feels optional and what users expect to happen next.
Examine
Ask for audience reactions tied to the task. Examine the findings and their evidence before turning a suspected issue into an implementation change.
Act
Use the findings to brief the next edit, then check the revised experience and validate it with real users.
Illustrative research brief. This is an example of how to use findings, not a completed customer study or a product screenshot.
Internal back-testing
UX testing. Measured against a human reference.
Sunzuin minutes75%
Human expert with AIin days34%
Research, 500+ humansin months100%
Sunzu internal back-testing. Research, 500+ humans is the reference set to 100%. This is a study-specific comparison, not an overall accuracy score or a guarantee for your study.
From our customers
Relied on by global experts
Director, Consumer Insights & Strategy
In just a couple of months, over 170 of our colleagues have run over 500 studies.
We’re now directly integrating audiences into many of our research, product and marketing workflows.
Director, Global Product and Design UX Research
In a market flooded with AI-generated noise, this was the first platform that genuinely earned our trust.
By grounding its models in research principles, behavioural science, and real customer data, it delivers insights that feel credible, transparent, and actionable.
It changed our perspective on what’s possible with AI in customer research.
Building a continuous UX review workflow?
Let’s scope the next step together.
Bring the user task and development workflow you want to review. Automated PR comments and a CI/CD review API are opportunities to discuss, rather than features included in the current MCP offer.
A few useful answers
Before your first study.
Depth interviews, concept tests and UX tests. Ad testing is available in the web app. MCP access depends on your account entitlements; the setup page contains your connection instructions.
No. The current offer on this page is research through MCP. Automated PR comments, a CI/CD review API and merge gates are proposed workflows to discuss with Sunzu, not available integrations advertised here.
Use a client compatible with Sunzu’s MCP connection instructions. The current public demo shows Claude listing available audiences; it does not demonstrate an automated pull-request review.
Sunzu creates audience characters grounded in customer evidence and enriched with a behavioural corpus. They are simulations. Use findings to explore directions and sharpen questions; lived experience, unexpected behaviour and consequential decisions still need research with real people.