Working Surfacelink · 21 Nov 2026
02
A heap of torn paper slips being shaken through a wide hand sieve into the hopper of a small weaving loom that turns out a neat patterned strip of cloth, with the person shaking the sieve seen from behind, the whole scene drawn with wide empty margins on every side.
Nielsen Norman Group

Can AI Generate Good Designs from Raw User Data?

Megan Chan and Rachel Banawa · 20 November 2026

How-to · artifacts · prototyping · Agent context files

Read the original
Takeaway

Nielsen Norman Group found that an AI model designed better interfaces from researchers' findings and recommendations than from raw user comments.

Summary

Megan Chan and Rachel Banawa of Nielsen Norman Group, a user-experience research and training firm, asked an AI model to design a flight-rebooking form and an online checkout flow. The AI model received either about 100 raw user comments or findings and recommendations that a researcher drew from the same comments. Designs built from the findings met more of a fixed scorecard's criteria.

Key points
  • Designs from findings and recommendations met 77.9 percent of the criteria, against 69.5 percent for designs from raw comments, a statistically significant gap.
  • No design built from raw comments told travellers whether a cancellation affected their loyalty miles, while every design built from findings did.
  • A human grader scored all 32 designs, and an AI model used as a second grader agreed on about 96 percent of scoring decisions.
  • The authors note that the comments were few and clean and the tasks familiar, and expect the gap to grow with messier research.
Implication

A design team that hands raw research to an AI design tool leaves the tool to decide which user problems matter. The authors recommend writing findings and recommendations with AI tools as one of their readers. Nielsen Norman Group also sells a course on this practice.

Plays
Design

Give an AI design tool synthesised research findings and recommendations, not raw user comments alone.

Design

Score AI-generated designs against a written checklist of criteria drawn from the research.

Derived by Working Surface from the article; more in the Playbook. Source line: Research artifacts now have two audiences: your teammates and the AI tools used in the design process.

Source issue

21 November 2026: as AI agents take on investigation, design and drafting, the human work moves to choosing what the agents read and checking what they produce before it reaches customers.