Skip to content
Article Designed Minds Jun 2026

Designed Minds: how designers are actually using AI in 2026

This Substack article by Eunji Jeong and the Designed Minds team summarizes findings from a June 11, 2026 workshop where designers shared real workflow examples rather than aspirational plans. 170 designers submitted a pre-workshop survey, 71 attended, and 31 completed a detailed workflow exercise in which they mapped where AI appeared — or didn’t — in a recent project.

The article is upfront about its limitations: the sample self-selected through the Designed Minds community and skews toward practitioners who already use AI. It is not a representative picture of the design profession as a whole. What it does offer is a close look at how designers who have committed to AI-assisted work actually distribute it across their process.

The hourglass pattern

The 31 workflow submissions, when analyzed together, showed a consistent shape: AI use was heaviest at the beginning and end of a project and much lighter in the middle.

At the start, designers used AI heavily for requirements gathering, synthesis, and research — both competitive analysis and inspiration gathering. At the end, they used it for prototyping and for preparing stakeholder materials such as presentations and summaries.

In the middle stages — sketching, early ideation, core visual design decisions — most designers worked manually. The article notes two possible explanations: that experienced designers treat these stages as the domain where craft and judgment matter most, or that current tools are simply not good enough there yet. The authors say the data cannot distinguish between the two.

Tools in the room

82% of workshop attendees use Claude regularly, making it the most common AI tool in the sample. Figma came in at 83%, which slightly edges out Claude, though the two are often used together. ChatGPT is used by 56% of respondents. Tools mentioned for specific tasks include Figma Make, Lovable, Cursor, Gamma, and Tactic.

70% of participants reported using AI every day. The workshop hosts were Parys G Khazaie, Michelle Yi Hsu, and Dhrumil Shah.

What to take from this

The article’s main practical advice is to start with synthesis. If a designer is not yet using AI regularly, requirements analysis and research synthesis are the entry points with the most accessible quality signals — it is relatively easy to judge whether a summary or a competitive overview is good. The prototyping end is similarly tractable. Sketching and core visual decisions are where most designers in this sample kept human judgment central, whether by choice or by necessity.

The piece is worth reading for any designer benchmarking their own AI use or trying to explain to a team where AI fits in a design process. The workshop format, which asked for real project examples rather than general opinions, makes the findings more grounded than survey data that relies on self-reporting about attitudes and intentions.