Is design roasting dead in the AI era?

Hey guys! Happy Saturday. Let’s talk about a topic independent designers (especially design studio owners) would be very interested: acquire clients through design roasts.

A few weeks ago I got on a call with a friend's company. Seed stage, one product designer wearing four hats. I'd pulled apart their website, social and product ahead of time — inconsistent type scale, three different color systems, visual variants that doesn’t look so polished.

I walked through it the way I'd walked through a hundred of these before.

The founder friend brought his only designer to the call as well. The designer got defensive. Not hostile, just tight: "That's an area we haven't paid attention to. Our effort's been on other things."

I left the call thinking I'd delivered the analysis badly. I hadn't. I'd delivered a good analysis to the wrong layer of the problem, and his sentence was the actual finding — I just didn't hear it as one.

What design roasting actually was

Since I started two years ago (before ai was this smart), redesigning a stranger's website and posting the before-and-after was the most reliable business development I had. Pick a site, rebuild it, post it side by side. Some of those people became clients. Many more watched.

I used to think it worked because it demonstrated taste.

It didn't. It worked because it cost something. Everyone scrolling past knew that redesign took a day, maybe half a day, and that I'd given it away to someone who hadn't asked and might never pay. The design was the visible part. The generosity was the message. That's a costly signal, and costly signals work precisely because they're expensive to fake.

Then the cost went to almost zero.

I can now produce that same redesign in under an hour. So can everyone else. And audiences adjusted faster than any of us expected — a feed full of AI-assisted redesigns doesn't read as generous, it reads as spam with good taste. The output looks identical to what worked in 2022. The signal underneath it is gone.

This is the part people miss when they say AI is making design worse. It isn't, particularly. It's making proof of effort worthless, and a lot of what we built our marketing on was proof of effort wearing a portfolio's clothes.

Here's the part I didn't see

My roasts used to be aimed at indie hackers. Solo founders who built their own site, weren't designers, never claimed to be. When I pointed out the type was inconsistent, visual flaws, there was no one to embarrass. They'd say thanks. Often they'd thank me and ship the fix right away.

Along the way my target moved to venture backed startups, and at that stage, the website has an author. A named employee with a LinkedIn and a manager who reads it. I would assume if I critique that homepage publicly, I'm not offering a helpful outside eye. I'm grading someone's work where their boss can see the grade.

"That's an area we haven't paid attention to."

That designer wasn't making excuses. He was telling me the inconsistency wasn't a taste failure at all — it was a resource allocation artifact. Four hats, no time, and the website came fifth. Nothing in my analysis addressed that, because my analysis assumed the problem was judgment.

Worse: he had the least power of anyone to fix it. He didn't choose the timeline, the headcount, or the category conventions he inherited. I aimed a critique at the one person in the building who couldn't act on it — and who, at a company I'd want to work with, is either my buyer or sits next to them.

You cannot pitch someone by first establishing that their work is the problem.

The uncomfortable truth

If execution is almost free, then giving away execution proves nothing about you. It just proves you have the same tools as everyone else.

So the work has to move upstream — into the reasoning, into knowing why a category looks the way it does, into being able to tell a client which conventions they have to keep and which ones they're following out of habit.

That's slower. I can't do it daily; it's more like weekly or monthly. It doesn't compress well, which is exactly why it's worth doing.

The old format let me demonstrate craft in a day. The new one requires actually understanding an industry before I'm allowed to have an opinion about it.

It is much harder for most designers. But, are you ready for the change and still willing to put in the efforts for some unknown results?

At least I will try.

Share my newsletter

Loving my content so far? I’d appreciate if you can share my newsletter to a friend 🙂