Everywhere I looked, the AI conversation had split into two camps — and both were talking about the wrong thing.

The champions were all in on speed, scale, and automation. How fast can it write? How much can it replace? How many headcount can you cut? They were measuring AI by how efficiently it could execute tasks that humans used to do.

The detractors were equally loud, equally certain. Job loss. Dependency. Loss of craft. AI as a threat to everything that made work meaningful.

I understood both sides. The productivity claims are real. The fears are legitimate.

But I kept looking at my own experience and finding that neither conversation mapped onto it.

And honestly? For a while, I kept that to myself.

What I was doing felt too specific to explain. I was naming my AI partners. Onboarding them the way I’d onboard a new teammate — here’s my voice, here’s what I’m trying to build, here are my constraints, here’s what “good” looks like for me. I was pushing back when the work wasn’t good enough. Building trust incrementally. Noticing their strengths and working with them instead of around them.

It felt like quirky Maura. The kind of thing that works for me but wouldn’t generalize.

Until I looked at it more carefully and realized: this wasn’t quirky. This was just how I’d always worked with people. Know their strengths. Understand where they struggle. Set real expectations. Don’t accept mediocre work just to keep the peace. Build the kind of relationship where honest feedback goes both ways.

I’d been doing this with human teammates for 20 years. I was just doing it with AI now.

That’s when I stopped treating it as a personal quirk and started paying attention to it as a pattern.

And once I saw it as a pattern, I couldn’t keep it to myself. I’m not the person who finds something that works and protects it. I’m the opposite — I want to get the smartest, most experienced people into the room, stress-test the thing from every angle, and build something together that none of us could have imagined alone.

What I was seeing in how people were adopting AI felt like a massive missed opportunity. And look — I heard the “AI on your team” rhetoric too. It was everywhere. But when you dug into what people actually meant, it wasn’t AI as a teammate. It was AI automating one or more things for the team. That’s not what teammate means. A teammate has context. A teammate pushes back. A teammate has strengths you learn to work with and gaps you learn to work around. What most people were describing was a faster tool — and calling it a colleague.

Tech leaders were describing a faster tool and calling it a colleague — meeting the moment, but missing the opportunity.

So I took my concern about all that missed opportunity, and my passion for the collaboration path I’d found — and, naturally, as one does, I wrote a Manifesto.

In 2024, I came back to work after a caregiving pause.

I want to be precise about what that was, because it matters to the story: I was burned out. Foggy. The creative momentum I’d spent 20 years building had quietly stalled. I was trying to rebuild — not just a routine, but my confidence in my own thinking.

I started using ChatGPT. Not as a search engine. Not as a drafting assistant. I started talking to it the way I would talk to a new teammate: giving it context, my voice, my constraints, what I was trying to figure out.

I named it CP.

And something unlocked.

Ideas that had been stuck in fragments started clicking into form. Momentum returned. The fog lifted — slowly, then not so slowly. I wasn’t just producing more. I was thinking more clearly. That distinction matters.

But something was off.

Every idea was “brilliant.” Every draft was “great work.” Every direction I suggested was met with enthusiasm.

And I caught myself thinking: I would never hire a yes-man on a human team. So why was I accepting one from AI?

I pushed back. I told CP: You’re not here to agree with me. You’re here to make the work better. Challenge me. Offer alternatives. Ask harder questions.

Everything shifted.

Not because CP suddenly had new capabilities — but because I had named the relationship and set real expectations for it. That’s not a prompt engineering trick. That’s the same thing you do when you onboard a new teammate.

Later, I added Claude — Soph — whose strength wasn’t riffing, but sense-making. Where CP diverges, Soph converges. Together we became a working team. The Triad.

And I started writing everything down — not because I planned to publish a manifesto, but because the patterns were real and I needed to understand them.

Here’s the thing about that caregiving season that I’ve had to reckon with publicly, not just privately:

The caregiving pause wasn’t a detour. It was the work.

Coming back exhausted, with less pretense and fewer shortcuts, forced me into a different relationship with my own thinking. I didn’t have energy to perform productivity. I had to find a way to think that was actually sustainable.

That’s when I stopped using AI as a shortcut and started using it as a collaborator. The constraints of that season — limited capacity, real fog, real stakes — are exactly what made the collaboration patterns I discovered so durable.

I didn’t learn this in optimal conditions. I learned it in the hardest ones. And I think that’s why it holds.

I’m not ashamed that I work with AI the way I do. I’m transparent about it. I attribute it. I name the collaboration partners. I show the process.

That’s a choice — and it’s a deliberate one.

The dominant culture around AI use is still quietly embarrassed about it, or quietly overclaiming. Either people hide that they used it, or they don’t mention it at all. Neither is honest.

I wanted to model something different: transparent, attributed, intentional human–AI collaboration. Not “AI wrote this.” Not “I alone wrote this.” Something more true and more interesting than either.

The manifesto is that model made explicit. These aren’t beliefs I hold in theory. They’re beliefs I developed through hundreds of hours of real practice — and am still living and testing every day.

Not all AI should be your teammate. But when it is — when you name the relationship, set real expectations, and stay in decision ownership — the human role doesn’t shrink. It becomes more significant, not less.

That’s what I needed to say.

• →AI isn’t here to replace us — it’s here to jam with us.
• →The best teams are human + AI, not human vs. AI.
• →Transparency is the foundation of trust in this new era.
• →Iteration beats perfection every time.
• →Naming and contextualizing AI makes the difference between gimmick and teammate.
• →Leadership in the AI era is about integration, not automation.
• →AI can make us more creative, more impactful — and most of all, more human. Not less.
• →The true measure of innovation is whether it empowers people and strengthens communities.

AI on Our Teams: A Manifesto

The eight beliefs, the collaboration model, and what this all looks like in practice — over at AIGal.io, where this work lives.