Your Company Already Said The Number That Wins Design Arguments

“Three of five users struggled with the signup flow.”

“We have an 80% onboarding abandonment rate.”

Same findings. But it feels like your team only listens to the second one.

You know the first line because user research is what you’re funded to collect. The second line, you probably assume, you don’t have access to.

That’s the part that’s changed with AI.

In Crucial Conversations: Tools for Talking When Stakes Are High, the authors talk about returning to one move to keep high-stakes conversations productive: when tempers flare, start with facts. Facts are the least controversial thing you can put on the table and the most persuasive.

It’s the first skill in their STATE model, and it’s what makes everything after it possible.

But facts on their own don’t cool a conversation down. What they do is let you re-establish the two conditions the book says dialogue depends on:

• Mutual Purpose

• Mutual Respect

By zooming out to the larger purpose, you re-ground the conversation, and productive discussion can follow. Tom Greever calls the same move “establishing common ground” in Articulating Design Decisions.

The problem is that the facts have to matter to both parties. And UX facts often don’t.

“We both want to fix the signup flow that three of five users struggled with, right?” That might not be true.

“We both want to fix an 80% onboarding abandonment rate, right?” That one is a business fact, which means your team already cares about it. You care about it too, just from the user’s side.

Grounding a conversation in shared purpose is one of the fundamental building blocks of handling high-stakes conversations well. But this is where designers say one specific thing:

“I don’t have access to data.”

That doesn’t have to be a limitation anymore.

Why does everyone keep showing up with the first line? Because it’s the only evidence anyone gives them room to collect.

There’s a set user research process to follow, and it’s something you’ve done at least once.

But the core of this newsletter, and of my philosophy, is data-informed design. Forming hypotheses, running experiments, A/B tests.

And while some of this stuff can be treated as a nice-to-have, an experimental mindset can easily solve the “I don’t have data” issue.

Even when deadlines are pushing in, everyone wants you to ship faster, and there isn’t a budget for it.

Evidence isn’t only the stuff you generate yourself. Some of it is already sitting in public.

I don’t necessarily trust AI to do design. But I do trust it to find facts with quoted sources, because it’s unlikely to hallucinate.

So imagine you don’t know why a project started. You’re on Project Neptune; it hasn’t been released yet, and you don’t really know much about it.

AI has never heard of Project Neptune either. It’s an internal codename, and it isn’t in any public document. Ask about it by name, and you’ll get nothing, or worse, you’ll get something invented.

So don’t ask AI to find Project Neptune. Ask it to find the problem that Project Neptune is meant to solve in boring places like shareholder meetings, press releases, and more.