How to Be an Indispensable UX for AI Designer

Some skills won’t just make you a better designer — they’ll make you indispensable. Like, “get pulled into every important meeting” indispensable. Like, “make yourself immune to reorgs” indispensable.

Will these skills 100% guarantee you survive the next budget guillotine? Of course not. (Anyone who promises job security in today’s AI-disrupted economy is either selling something or is in HR.) But mastering these skills will radically increase your odds of survival, allow you to contribute at a higher level, and make your work way more fun in the process.

Yes, you can learn them. Yes, I’ll help you. And yes, there’s a link at the bottom if you want to accelerate.

Let’s go!

1. Bias for Action (a.k.a. Vibe Code or Die Trying)

Stop waiting for requirements. Ask better questions:
❌ “What am I supposed to do?”
✅ “What can I build that’s actually useful?”

When you combine a basic understanding of user workflows with even a modicum of technical curiosity, you become lethal. You can use LLMs to generate ugly-but-functional working demos in under a day. Suddenly, you’re not just “designing screens” — you’re designing systems.

This isn’t just speed. It’s leverage.

2. Learn to Dance (Just Like the Bees Do)

When bees find a great flower patch, they do a little dance to convince the rest of the hive to come with them. Your job is the same: demo (dance) your way to influence.

Your AI prototype needs to tell a story. It needs to be compelling. It needs to make the PM lean forward and say, “Wait… can it really do that?”
And yes, it should run. Real model. Real data. Real outputs. Stop building static crap.

Want buy-in? Don’t wait for permission. Put on your wings and start dancing.

3. Make Soup from a Stone

There’s an old fable about a soldier who tricks a village into sharing their ingredients by starting with just a stone in a pot… So they create and eat delicious soup. Together.

UX is that stone.

You bring people to the table. You create a shared vision, a working prototype, a single point of traction. This isn’t about mockups — it’s about momentum. You build the thing that gets everyone talking.

Today, UX isn't crayons in the corner. UX is the catalyst.

4. Create the Data

AI isn’t magic. It runs on data — and increasingly, that data comes from use cases.

Guess who owns the use cases?

(Pssst… You do, of course.)

So stop outsourcing your brain to “the data team.” Own the RAG pipeline. Define the inputs. Validate the outputs. Partner with your data scientist and your SME, and start building systems that learn.

This is the UX-Data-AI loop. Live it.

5. Be Ruthlessly Iterative

Waterfall is dead. Agile theater is worse. The old game of “PM writes PRD → UX makes mockup → Engineers laugh and build something else” is a shambling zombie — and it’s still eating people’s brains (and time.)

Instead, do this:
Build an end-to-end MVP with three ingredients:

• Basic UX (even ugly)

• Actual data

• A functioning AI model

Ship that. Then layer. Iterate. Evolve. MVP should mean Minimal Viable Product, not Most Vaguely Promised.

6. Extreme Ownership (No More T-Rex Hands 🦖)

If the product’s broken, guess whose problem it is?

Yours.

As a UX for AI designer, you own the outcome. You don’t get to say “well, the devs didn’t…” or “QA missed…” or “PM never told me…” No.

You saw it. You fix it. Or you escalate.

Or you stand in front of it like a Roman damn shield.

Own the product. Full stop.

A Day in the Life of a UX for AI Designer

Your day should look something like this:

• Start with a cross-functional project standup. What did you do? What are you doing? Any blockers? (60 seconds. Take everything else offline.)

• Meet 1–2 customers. Co-create with a doc cam and a pencil. Explore data flows, prompt formatting, edge cases.

• Sync with your data scientist + PM. Define how the AI learns. Trace it from user intent → data → output.

• Work a Jira ticket. AI failed? Sit down with your SME + engineer. Update the RAG. Test. Validate. Close the damn ticket.

• Vibe-code your next demo. From your favorite coffee shop. With an LLM as your rubber duck. Real inputs. Real models.

Notice what’s not in that list?

Figma.

No re-recreating a damn data table cell-by-cell. For the 1000th time.
No copying LLM responses into high-fidelity mockups.
No pixel pushing.

If your design job still revolves around Figma, it’s time to ask: Are you building a product or just standing on the deck of the Titanic, busily rearranging the chairs and waiting for the inevitable?

Figma is Fine. Until It’s the Titanic.

Don’t get me wrong — Figma’s a decent tool for documentation for remote teams/contractors, some high-fidelity usability testing, and visual polish. But if you’re still clinging to it like it’s your lifeboat, you’re going down with the ship.

Modern UX for AI isn’t in the pixels. It’s in the flows.
It’s in React pages running real data from real APIs.
It’s in shipping, testing, and tuning live AI behavior.

👋 Want help becoming indispensable?

I’m opening up 3 private coaching spots for UXers who want to break through blockers, transition into AI leadership, file patents, write that book, or just finally build something real.

If that sounds like you, email me at Greg [at] UXforAI.com — use “Coaching” in the subject line, and tell me why you want one of the three slots.

No fluff. No public forms. Just a real conversation with someone who’s done it before (and written a few books about it).

Let’s build the future — together.
— Greg