In a recent LinkedIn post, Nasir Uddin shares a practical, AI-assisted workflow for analyzing UI design, moving beyond subjective gut feelings to identify concrete usability issues. Uddin highlights a common designer’s dilemma: a visually polished UI that still fails to convert or convert users effectively.
According to Uddin, the key lies in leveraging AI tools like Gemini 3 Pro, but with a structured approach. He emphasizes that the quality of AI feedback is directly tied to the input provided. “Garbage input always gives garbage feedback,” Uddin states, underscoring the importance of using real product screens from actual user flows, such as login, onboarding, or checkout processes, rather than idealized concepts or Dribbble shots.
The Importance of Clean and Intentional Input
Uddin outlines a step-by-step process designed to maximize the effectiveness of AI analysis. The initial steps focus on preparing the design assets for the AI. This involves exporting actual product screens, ensuring labels are visible and that the versions are final and connected within a flow. Uddin advises designers to “clean the noise” by removing unused variations or old experiments, treating the preparation phase as a meticulous review process.
“Keep labels visible. Avoid half-finished versions. Clarity matters more than quantity.”
This emphasis on clarity and intention, as Uddin points out, ensures that the AI receives the most relevant data, leading to more accurate and actionable insights. The goal is to feed the AI a representative sample of the user experience, not a curated or incomplete picture.
Crafting Effective Prompts for Actionable Insights
A critical phase in Uddin’s workflow involves asking the right questions. He cautions against simply uploading screens and hoping for the best. Instead, Uddin suggests being specific with prompts to guide the AI’s analysis. Examples he provides include asking, “What usability issues do you see?” or “Where might users get confused?”
As Uddin notes, “Good prompts unlock good insights.” This interactive approach helps to move the AI from a generic analysis tool to a targeted problem-solver. By framing specific queries, designers can elicit feedback that directly addresses potential pain points in the user journey.
Identifying Patterns and Structuring Feedback
Uddin further stresses the need to focus on patterns rather than isolated comments. “One comment can be ignored. Repeated feedback is a signal,” he argues. Designers should look for recurring themes such as friction points, areas of confusion, or hierarchy problems. This focus on consistent signals helps to prioritize issues that have the most significant impact on the user experience.
“Look for: Friction points. Confusion. Hierarchy problems. Patterns matter more than details.”
To make the AI’s output digestible and actionable, Uddin recommends structuring the feedback. He suggests asking the AI to provide summaries, prioritize issues, or present them in simple tables or lists. This structured output transforms raw data into a clear roadmap for design improvements.
Conclusion: AI as a Design Partner
Ultimately, Uddin positions AI not as a replacement for design intuition but as a powerful partner. By following this structured workflow, designers can move beyond guesswork and address real problems with confidence, leading to clearer interfaces, refined flows, and stronger visual hierarchy. He concludes by urging designers to consider the limitations of relying solely on “gut feeling” and to embrace AI-assisted analysis for enhanced clarity. “If you’re still reviewing your UI only by gut feeling, you’re leaving clarity on the table,” Uddin warns.
📝 About This Content
This article is based on insights shared by Nasir Uddin on LinkedIn.
📅 Originally posted on February 25, 2026 | View original post on LinkedIn →