Case Study 2026

UX Risk Review

AI-powered Figma plugin that helps Product Designers systematically review user flows before handoff.

Role
Product Designer
Scope
Discovery, Strategy, UX/UI
Users
Product & UX Designers
Category
AI for Product Design
Platform
Figma Plugin
UX Risk Review plugin screens showing the review workflow

The Missing Step

Before handoff, every screen has been designed.
Not every scenario has been reviewed.

Loading states, empty states, error handling and recovery flows are still often overlooked. Not because designers don't know they exist, but because there isn't a structured way to review a flow before handing it over to development.

That problem became the starting point for this research.

01 · Flow Logic

Review navigation, transitions and decision points.

02 · Missing States

Check loading, empty and edge-case scenarios.

03 · Failure Handling

Evaluate errors, recovery and fallback flows.

04 · Trust & Clarity

Identify issues that may confuse users.

05 · Scenario Exploration

Review alternative paths beyond the happy flow.

02

Research

The research explored how Product Designers review user flows before handoff and where AI could provide meaningful support.

Qualitative interviews were combined with secondary research, community discussions
and a review of existing AI design tools.

Research Methods
01

Interviews

Conversations with Product Designers about review workflows, handoff practices and overlooked scenarios.

02

Secondary Research

Industry articles, UX resources and best practices around design reviews and edge cases.

03

Community Insights

Discussions from design communities highlighting recurring review challenges and blind spots.

04

Competitor Analysis

Analysis of AI design assistants to understand current approaches, limitations and opportunities.

Research Insights

01

Reviews rely on memory

Without a structured process, important scenarios are easy to overlook.

02

Risks are discovered too late

Many UX issues surface during development or QA, when changes become more expensive.

03

Reviews need structure

AI creates the most value by guiding the review process, not replacing design decisions.

03

The Review Experience

Prepare the Review
UX Risk Review plugin screen for selecting frames, project context and review areas

Flow Detection. Automatically recognizes the selected user flow.

Project Context. Grounds the review in product-specific information.

Review Areas. Choose what to validate before starting the session.

UX Risk Review plugin preparing and mapping a selected prototype flow

Flow Analysis. Maps selected screens, prototype connections and project context.

Review Summary. Shows what was analysed and how findings are distributed across review areas.

Clear Expectations. Previews the scope and importance of findings before the session begins.

UX Risk Review summary of analysed screens, connections and review areas

Flow Analysis. Maps selected screens, prototype connections and project context.

Review Summary. Shows what was analysed and how findings are distributed across review areas.

Clear Expectations. Previews the scope and importance of findings before the session begins.

Review Session
UX Risk Review session with findings grouped by review area and decision controls

Structured Review. Every finding is grouped by review area and presented with clear context.

AI Discussion. Each observation can be discussed with an AI agent to explore alternatives, ask follow-up questions or challenge the suggestion.

Decision Tracking. Mark findings as reviewed, not applicable or requiring discussion to keep the review process transparent.

Senior Design Perspective. The agent responds as an experienced Product Designer, explaining not only what might be wrong, but why it matters.

Instead of providing final answers, the agent highlights trade-offs, asks clarifying questions and helps teams make informed design decisions.

UX Risk Review discussion agent offering a senior design perspective
UX Risk Review scenario questions for paths beyond the happy path

Review beyond the happy path. Highlights user scenarios that could otherwise be missed during review.

Review Summary
UX Risk Review completion state with remaining decisions

Review Summary. Generate a structured summary that captures decisions, unresolved items and review progress.

Share Anywhere. Copy the summary into Slack, Linear, Jira or any team chat or generate a Figma frame that stays alongside the design files.

One Source of Truth. Keep review decisions, discussion points and unresolved questions connected to the design throughout development.

Decision Overview. Summarizes reviewed items, scenarios and discussion points in one place.

Ready for Handoff. The review is complete once every discussion point has been resolved.

UX Risk Review completion state ready for handoff
Generated UX review summary frame for sharing with the product team
04

Built for Product Teams

01

Bilingual Interface

Switch between English and Ukrainian at any time.

02

Project Context

Grounds every review in product-specific information.

03

Discussion Agent

Supports design decisions through discussion.

04

Team-Ready Output

Generate summaries or Figma review frames.

05

Session Memory

Continue working on the current review or start a new session without losing previous context.

05

Beyond the MVP

Several ideas were intentionally left out of the first version to keep the core review experience focused and lightweight.

Team Knowledge Base

Build a shared library of recurring UX risks, review patterns and decisions across projects.

Integrations

Send review findings and unresolved discussion points directly to Jira, Linear or Slack.

Design System Awareness

Use existing components, patterns and guidelines as additional context during the review.

Review Templates

Save reusable review configurations for common flows such as onboarding, checkout or authentication.

06 Reflection

The hardest part wasn't adding AI to the workflow.
It was defining where the designer takes over.

Instead of trying to replace design decisions, the goal became helping designers make better ones.

The biggest challenge was finding the boundary between guidance and autonomy. The AI needed to identify risks, explain their impact and ask the right questions without becoming another stakeholder making product decisions.

That boundary ultimately shaped the entire product. Every recommendation is designed to support critical thinking rather than replace it, leaving the final decision with the designer.

Primary Insight Good AI doesn't replace judgment.
It strengthens it.
Core Decision Recommend. Explain. Never decide.
Design Principle Supporting designer decisions.
Next Project

Sky Alert

A mobile app that helps people quickly understand an air raid alert and reach the nearest shelter.

View Case Study
Sky Alert case study preview