Example · Product / UX Designer
48% of your week as a product and UX designer is still yours.
AI now produces screens in minutes. Judging which ones are right stays with you.
Make my planA typical week at the middle pace. Your plan uses your own tasks.
Is UX design dying because of AI?
In a typical week at the middle pace, 35% of this work needs your judgment and 13% is work you do using AI. AI could do about 52%. Point a design tool like Figma Make at a settings flow with your library connected, and it gets most of the way in minutes. It cannot tell which parts are wrong.
Read more
Your job now works the same way. Producing screens is getting cheap, but knowing which screens are wrong still takes a person. About 53% of your week is the first kind of work. A third is the second kind, most of it work you are allowed to do but were not hired to do.
Two things come before the numbers. First, the hard year for designers already happened, and AI did not cause it. Job postings for UX designers fell 71% from 2022 to 2023 as companies unwound a hiring bubble (Jakob Nielsen, UX Tigers, September 2023); many design teams shrank in that correction and never refilled those seats. Second, the parts of your job that hold are framing the problem with a PM, sitting with a customer, defending a decision in a room. That last one, the stakeholder review where a senior leader redesigns your flow on the call, is hard for many designers. This report shows you how to walk in holding evidence they cannot redesign.
The week, task by task
Who does each hour of a typical 40-hour week.
See every task
Needs your judgment.
- Framing the problem with a PM: discovery, scoping, "what are we solving"8%
- Critique, stakeholder reviews, negotiating scope7%
- Moderating usability sessions and customer interviews6%
- Standup, planning, design team sync6%
- + 2 smaller tasks8%
AI does the slow parts. You still own it.
- Research ops and synthesis: recruiting, scheduling, notes to themes6%
- Design QA of built UI and engineers' build-time questions6%
AI could do most of it.
- Hi-fi screens from an agreed direction, on the design system18%
- Wireframes and exploring alternatives from a brief10%
- Design-system maintenance: variants, tokens, component docs8%
- Clickable prototypes for review and testing7%
- + 2 smaller tasks10%
Why some tasks still need a person
Turning a PM's brief into the actual problem
What to build lives in six months of support tickets, two lost deals and a PM's half-formed hunch, and none of it is in the Figma file or the codebase. That is context a model cannot read, and it is the step that happens before any tool can start.
Sitting in a room with a customer and watching them fail
Live moderation needs presence, follow-up questions, and the judgment to notice what they did not say. This is checking work, because you are the one who finds out whether the design is wrong.
Knowing why the design system has two button components
This rests on context nobody wrote down. The second one exists because of a customer contract nobody wrote down. A model reading the library sees two buttons, but you see a decision.
Answering the engineer's "what happens when the list is empty?" on build day
The answer depends on a customer conversation from months ago and what the PM promised in a quarterly business review. It is cheap to ask you and expensive to reconstruct.
Saying "this shipped screen does not match, and here is why it matters"
This duty comes from the product's contracts. Enterprise customers who ask for a WCAG conformance report need a named person's judgment behind it.
Good to know
What is hard about this field, and what you can do about it.
The hard truth
You will be judged on business results, more than screens.
Leaders now want proof that design moved the business. Polished screens are getting cheap to make, so they no longer make the case for you.
Sources (2)
- Leaders demanded clearer proof design affected the bottom line, and UX often struggled to show it. Nielsen Norman Group, State of UX 2026, 2026-01
- Entry-level UX roles remain scarce, with more aspiring designers than open roles. Nielsen Norman Group, State of UX 2026, 2026-01
What's next for a UX designer?
Plan A is the best move inside the work you do now.
Grow in the job you have
The main focus. Most weekly sessions build toward it.
- Each week
- 4 sessions
- Ready in
- About 4 to 6 months
Why it fits
Why you. If you have moderated research rounds, own the flows customers complain about most, and want research that changes the roadmap instead of decorating it, this is the move. Standalone UX research roles are close to gone (922 postings in the June 2026 tracker against 16,410 for product design), so research now happens inside design or not at all.
The gap. You have no standing cadence, no participant pipeline of your own (PMs often hold customer access) and no habit of turning findings into a one-page memo a VP reads.
The trade-off. Your visible output drops for a quarter, at a company that may count shipped screens. A PM may read weekly customer calls as a turf move. You need the Head of Design to name this as your scope early, which is week 3.
2 other moves that fit
- Run the agent instead of racing itReady in about 2 to 3 months
- Hold the accessibility and design-system governance lineReady in about 3 to 4 months
What's a plan B for product designers?
Plan B is meant to be different from this job. It was picked because skills from this role carry over.
Get ready for a new kind of job
A second option. A newer job AI is creating that these skills already fit.
Watching customers fail in a session gives you the empathy and flow thinking that designing an AI agent's conversations calls for.
- Each week
- 1 session
- Outlook
- Hiring now
- Pay
- $82K to $200K a yearThree US roles at Cresta, Checkr (closed) and Decagon, latest Oct 2026; the top end asks for 5+ years.
2 other options
- AI product managerYou already turn a PM's brief into the problem worth solving and defend a direction to leadership, which is half of owning an AI feature.
- AI evaluation specialistSpotting where a shipped screen drifts from the design and explaining why it matters translates into judging AI output quality.
In your own plan you can switch to any of the 47 roles.
The 13 weeks
Small steps that add up. One plan for both: most sessions build Plan A, one a week builds Plan B.
By week 13 you will have
AFor Plan A
- Nine customer interviews with clips
- Three decision memos
- A published case study
BAnd for Plan B
- Scored test results and a revised flow
- A conversation design case study for your portfolio
- 1Measure your week
- 2Get customer access and set rules
- 3Ask for research scope
- 4Write your first memo
You finish withAn accessibility audit with a failure count
- 5Test the rules file
- 6Play a clip in review
- 7Back evidence with a number
- 8Share the rules file
You finish withAn opportunity tree, a map of customer needs, for one flow
- 9Fix and govern the system
- 10Count your results
- 11Write memo and case study
- 12Make your ask
- 13Publish and look back
You finish withA one-page review and your next plan
Three courses from the plan
- Just Enough Research, 2024 editionMule Books (Erika Hall) · Paperback or DRM-free ebook; price set by the bookseller
Yes, week 2. It is the one book in this plan you should read cover to cover.
- Continuous Discovery HabitsProduct Talk (Teresa Torres) · Paperback or Kindle; price set by the bookseller
Yes, week 8. Chapters 1–5 are enough.
- Product Discovery Fundamentals (live cohort)Product Talk Academy · $1,795 per seat
Don't buy this yet. Do nine interviews first; then decide if you need a course.
Checked 2026-09-22.
The fuller picture
Risks beyond AI: moderatePost-bubble hiring risk
The 2022–23 correction was the most important event in this market and it predates the tools. A small software company that cut design once can cut it again on a funding event. This report models AI exposure, but it cannot predict your CFO's choices.
Where this could be wrong
- The shares are estimates until week 1. If your log shows hi-fi screens are a much bigger slice, more of your week is work AI can do, and running the agent becomes urgent this quarter.
- We may have scored wireframing too high. Our score assumes a good brief. If your PMs' briefs are usually bad, exploration is problem framing in wireframe form, and belongs with the work that lasts. That means more of your work lasts, and it moves the research move up.
- The 2022–23 correction may not be over for your company. A second cut can follow a funding event, and nothing in this report can see that coming.
- AI-moderated interview tools may get trusted faster than we think. If customers accept an AI interviewer within two years, AI gets much better at moderation and the research move loses part of its floor. The problem-framing half still holds.
Also matched: ux designer, product designer, ui designer. Research checked 2026-09-22. Page updated .