Overview
Wix Prizma is an internal AI-powered platform that helps teams run early usability tests for product validation. It combines behavioral data from multiple internal sources with AI-generated user simulations to identify usability issues, uncover potential bugs and edge cases, validate concepts, and generate actionable UX insights before features reach development or production.
The Problem
Traditional usability testing is often slow, expensive, and difficult to run in the early stages of product development. Teams need a faster way to validate ideas, uncover usability issues, and gather meaningful feedback before investing engineering time in development.
The Goal
An internal AI platform that helps teams validate features before development. Combining FullStory, BI events, Sled and other internal data, Prizma uses LLM-based simulations of real users to surface friction and turn complex data into insights for designers, PMs and researchers.
Research
Key insights
64%
Of usability issues are discovered only after development begins
58%
Of teams report costly design iterations due to late user validation
79%
Believe AI-assisted testing could accelerate product decisions
Primary Users
Key user groups
Product Designers
Need fast usability validation before development.
Product Managers
Want confidence before investing engineering resources.
UX Researchers
Need scalable testing without recruiting participants.
Process
How I solved it
Making early product validation faster and smarter through AI-generated user simulations, helping teams uncover usability issues before development.
Integrating all data sources
Aggregates data from FullStory, BI events, Sled and other internal Wix tools into one comprehensive behavioral dataset.
Simulating real user behavior
Uses the combined dataset to train an AI model that simulates how real users interact with product features.
Real-time, simple analytics
A real-time dashboard that visualizes and summarizes how simulated users interact with the tested feature.
Actionable recommendations
Turns behavioral data into meaningful insights, including tailored recommendations and warnings that help designers improve the experience.
The value of an AI product depends on how clearly it supports human judgment.
Process
Wix Prizma flow
I mapped the entire user journey from uploading a design to reviewing AI-generated insights. The flow outlines each key step, decision and interaction, ensuring a structured validation process that helps teams evaluate products with confidence before development.
Featured Screens
From uploading a design to reviewing AI insights.
Final Design Solution
A simple, focused validation process.
From uploading a design to reviewing AI-generated insights, every step helps teams spot friction, understand user behavior, and improve the product experience before investing development time.
Reflection
What I learned
Designing Wix Prizma reinforced that the value of an AI product depends on how clearly it supports human judgment. The most important challenge was not presenting more data, but creating the right hierarchy, context and guidance, so teams could understand the results, recognize uncertainty, and confidently decide what to do next.
