Published in Bootcamp, a UX Collective publication on Medium.

Between Figma Make AI, Cursor and all the other AI UX stuff floating around, I wanted to see two things for myself:

1. Is this actually a threat to UX designers?

2. How can I use it without slowly hating my craft?

Spoiler: my job is still here. But parts of it have definitely changed.

The fear phase: “AI will take UX jobs”

Like most people, I had that familiar anxiety: UX is not that special, AI can generate interfaces in seconds, and maybe in a year we will all be retired against our will.

So I did the experiment properly. I paid for Cursor, tried Figma Make AI, and plugged them into my real world: complex B2B enterprise tools, not shiny e‑commerce card grids.

Very quickly it became clear:
AI can absolutely fake a nice looking UI.
But it cannot fake context.

Where UI AI falls flat: B2B is not a landing page

For the kind of network planning tools and enterprise products I work on, the UI is just the visible tip of a very complicated iceberg.

AI can do:

• Generic layouts

• Decent buttons, cards, filters

• Something that looks like an app

What it cannot do (yet):

• Understand the mental model of a network planner

• Prioritise information in a way that matches real workflows

• Reflect all the weird edge cases that product teams have learned through painful user interviews

• Translate messy business rules into clean, usable interaction patterns

Even with good prompts, I found I was using maybe 20–30 percent of what AI generated. The rest still required what humans are annoyingly good at: talking to other humans, understanding context, negotiating constraints, iterating based on real feedback.

AI gave me a starting point.
But it did not give me a product.

AI as a brainstorming buddy, not a replacement

Where AI actually shines is brainstorming, especially for people who are not designers.

For example, it is helpful when:

• A product owner has a vague idea and no way to visualise it

• A business stakeholder says “I just need a dashboard” and you need to tease out what that might look like

• An engineer wants to propose a flow but is more comfortable in code than in Figma

AI is great for:

• Trying different filter layouts

• Playing with variations of a card or table

• Getting a few quick directions for one part of a page

But: the overall information architecture, interaction model and hierarchy still need a designer. AI can remix patterns. It does not decide what actually matters for the user or how the whole system holds together.

Solo designer? AI becomes your unofficial team mate

If you are the only designer on a product team, AI is weirdly helpful as a fake colleague.

You can:

• Ask it to propose alternative flows you have not considered

• Get quick critiques on your copy or hierarchy

• Generate a few “what if we tried this layout instead” options to unblock yourself

It is not a replacement for proper collaboration. But if you lack a design team to bounce ideas off, AI can at least bring new perspectives into the room.

That said, it still does not:

• Decide scope

• Frame the problem for your product team

• Run user interviews

• Facilitate ideation workshops

• Align stakeholders

Those parts are still 100 percent human. And that is probably where the real value of UX is moving anyway.

What AI is definitely not: a PO or engineer shortcut to “instant product”

One thing that worries me a bit is this fantasy:
“A frontend engineer or PO just prompts Figma AI, gets a design, ships whatever comes out. Done.”

That is exactly how you create:

• Inconsistent experiences

• Features nobody understands

• “It looks pretty but feels wrong” products

Treating AI output as ready to ship design is like treating lorem ipsum as final content. It might be fine for a prototype. It is not fine for production.

AI is a tool in the UX process, not a bypass of it.

The interesting part: Figma Make AI + design systems + code

Now to the part that genuinely excites me.

Imagine this:

• Your company’s design system components are fully integrated with Figma Make AI

• The AI also knows your default tech stack, say React with your internal component library

• You prompt Figma, it generates a flow using your actual components

• And alongside the design, it spits out production‑ready code

Suddenly, handover looks very different.

It raises some big questions:

• Do designers start delivering “design + code” as one package?

• Does frontend engineering on standard UI become more like reviewing and refining, rather than building from scratch?

• If the code is ready, do product owners lose the classic excuse of “we do not have time to tackle this design debt now”? Because the work is mostly done and just needs review and merge.

I am very curious to see how this part evolves.
If done right, it could significantly reduce design debt and make it harder to ignore UX in delivery.

Where Figma Make AI actually works for B2B

Right now, for complex enterprise products, Figma Make AI seems most useful in a very specific way:

1. You still create the wireframes yourself.

2. You define the flows, information architecture and key states.

3. Then you ask Figma AI to:

   • Apply your design system

   • Fill in realistic content

   • Generate variants and edge states

   • Help with the “make it look real” step

In other words, designers still think.
AI helps polish.

That feels like a good balance for the current state of the tools.

The quiet revolution: research, sensemaking and insight synthesis

The biggest productivity gain for me was not UI at all. It was research.

Using AI for:

• Affinity mapping

• Clustering qualitative data

• Drafting initial insight summaries

• Suggesting themes across interview notes

This shifted tasks that used to take 1–2 weeks down to something like 4–8 hours. That is huge.

Of course, you still need a human to:

• Check for hallucinations and wrong groupings

• Correct misinterpretations of user quotes

• Turn “raw insights” into a story stakeholders will actually care about

• Build a compelling, context‑aware slide deck

AI speeds up the sorting.
It does not replace the storytelling.

So… should designers be worried?

At the beginning, yes, it felt scary.
Now, it feels more like this:

Every designer will need to adopt AI ways of working.
Not because AI will replace UX, but because designers who use AI well will outpace those who do not.

AI can:

• Accelerate brainstorming

• Automate repetitive research tasks

• Help generate first drafts of UI

• Assist with documentation and summarising

Designers need to:

• Own the problem framing

• Bring real user insight into the room

• Challenge AI’s “pretty but wrong” suggestions

• Focus more on strategy, systems and long term product value

The job shifts from “I push pixels” to “I design systems, narratives and decisions, with AI as a power tool.”

Published in Bootcamp, a UX Collective publication on Medium.