🧠 When AI Is the One Who Knows the Brief

How I Accidentally Confused My Team by Working Smarter (Not Lazier)
by Brady Starr

Designing the Proposal with AI: When your collaborator never sleeps, but sometimes makes things awkward.

Let me set the scene.

It’s a regular Tuesday morning. I’m in a standup with my business partner Olya and our teammate Jackie. We’re midway through responding to a hefty government RFP — one of those 50-page beasts that includes everything from accessibility compliance to SEO integration to detailed audience segmentation. The kind of project that’s big, important, and very much all-hands-on-deck.

We’re reviewing progress. I throw out a question:

“Are we like… 80–85% there on the proposal?”

There’s a beat. Then nods. Yeah, that sounds about right.

Then I say:

“Cool. Because, uh… I haven’t actually done anything yet.”

🫠

The moment hangs. I’ve just dropped what sounds like: I haven’t contributed. Like I’m cruising while everyone else is grinding.

But that’s not what I meant. Not even close.

🔍 A Process That Looked Like Nothing

What I was trying — and failing — to explain is that I had been working. A lot, actually. Just… differently.

From the beginning, instead of reading through the entire RFP myself, I fed it into ChatGPT. I asked it:

• “Where should we start?”
• “What are the key goals here?”
• “What should go in Attachment B?”
• “What roles are needed on this team?”
• “Can you rephrase this in plain language?”
• “What else are they expecting?”

Over and over, I queried, refined, structured. Piece by piece, it helped me draft a skeleton for every section, remember what belonged where, and keep track of overlapping requirements. I wasn’t just using AI to spit out content — I was using it to hold the architecture in its head so I didn’t have to.

That’s not nothing. That’s everything for someone like me — ADHD, juggling 14 browser tabs of mental noise at all times, trying to hold it together while leading a team and growing a business.

But in the moment, what came out of my mouth sounded like a confession of laziness.

🤖 Why I Let AI Lead

This wasn’t my first time doing things this way. During our PropertyWorx project, I built an AI assistant we called Aaron Interpreter — a digital teammate trained on 14+ dense documents from a founder who’d brain-dumped his entire strategy into Google Docs and called it a brief.

Read: 🧠 When the AI Writes the Brief: What It’s Like to Build with a Client Like Aaron

Aaron read everything. He remembered, synthesized, and quoted back better than any human could’ve. It was wild. It was useful. It was necessary.

So I tried the same thing here. The truth is: I didn’t read the full RFP. I didn’t need to. Not because I was skipping steps — but because I was using a system that could read it faster, organize it better, and surface the relevant information when I needed it.

It let me show up with clarity and creative energy instead of overwhelm.

It worked.

But I didn’t explain that. I just said: “I haven’t done anything.”

💬 The Human Cost of a Robotic Sentence

Olya, rightfully, pushed back. From her perspective, this felt like I was distancing myself from the work — removing the human from the process. And in UX, that’s sacrilege.

Jackie, who’s been pouring hours into the proposal, may have interpreted my words as dismissive. And I get that.

I later explained everything. My wife, when I debriefed the whole situation with her, had the same reaction: confusion, then understanding, then appreciation. Because it is confusing. The way I work now doesn’t always look like work.

But that’s the point.

🧭 This Is the New Brief

In the age of AI, we’re not always working from neat pitch decks or tidy PDFs. We’re working from:

• Brain dumps
• Loosely structured Google Docs
• ChatGPT-generated fragments
• Past proposals and overlapping expectations

The new “brief” is a constellation of documents, thoughts, and interpretations — not a singular source of truth.

And AI is the only thing that can hold the constellation in its head.

That’s what I was trying to do here. Not avoid the work. Not outsource the thinking. But augment my capacity to think clearly in the middle of chaos. To be a better strategist, designer, and leader — not a lesser one.

🧵 What I Learned (and What I’d Do Differently)

1. Explain your process early
   If you’re working differently — especially with AI — say so upfront. Don’t let people interpret silence as absence.
2. Your method is valid — but communication is key
   If your teammates are working their butts off and you say “I haven’t done anything,” it doesn’t matter how brilliant your backend process was. You’ve still stepped on their toes.
3. AI doesn’t make you less human — but it can make you sound that way
   Learn to translate your AI-powered methods into emotional language that still honors the human effort in the room.
4. This work counts
   Working with AI to structure, synthesize, and guide is still UX work. It’s invisible sometimes, but no less important.

🛠️ Final Thought: Designers, Use the Tools That Work for You

I’m not saying every designer should stop reading documents. I’m saying: if you’re wired in a way where linear reading drains your energy, but querying an AI assistant sparks insight — go for it. Just don’t forget to bring your team along for the ride.

This is the edge we’re all walking right now. The edge between old workflows and new capabilities. Between clarity and confusion. Between silence and contribution.

And next time? I’ll just say:

“I’ve been working with our AI assistant to get this structured — I’ve got drafts and direction queued up. Want to review it together?”

That feels a little more human.

—
Brady Starr
Co-Founder, Brady UX
bradyux.com