I built my first AI skill ten years ago

A few weeks ago, I wrote about how to stop being the PM who always says no.

Today I want to flip that around.

Because sometimes the harder thing isn't saying no.

It's saying yes when you have absolutely no idea what you're doing.

Today I want to share a personal story of when I built my first AI Skill before “AI PM” was a thing (let’s be real, it’s still not a real thing IMO).

The pitch that went sideways

In early 2015, I was a few months into my role at Amazon.

Still learning how things worked. Still trying to figure out who to CC on what email.

At the time, Amazon was quietly building something that would change how people interacted with technology.

The Alexa Skills Kit was a developer framework that would let anyone build voice-powered experiences on Alexa. It was in development but hadn't launched yet.

Internally, there was a lot of energy around it.

Someone on my team pushed me to pitch an idea to two senior executives. One of them was a VP named Greg.

I didn't feel ready.

But the meeting was on the calendar. So I walked in.

I remember sliding around printouts of a doc I wrote, ready to put a landing page mockup on screen, trying to pitch a concept that was half-formed at best.

Greg looked at me and said "What are we even doing here? We're not ready for this."

And he was right.

The idea wasn't baked. The timing wasn't right.

I'd been pushed into the room before I had something worth pitching.

It was one of those moments where you just want the floor to swallow you whole.

I didn’t even know Greg, yet somehow I was this new PM who’d been asked to put together an online experience for announcing what would be like a “Come build custom GPTs for ChatGPT” equivalent for OpenAI.

I was in way over my head.

What I got wrong

Let me be honest about this.

I walked into that room because someone told me I should. Not because I had a clear story to tell.

I didn't connect my idea to a problem the VPs cared about.

I didn't anchor it in data or customer insight.

I just... put an idea together.

The opportunity itself wasn't the problem. The opportunity was real.

Voice was about to explode. Amazon was months away from opening up the Alexa Skills Kit to external developers.

Within a year, there would be nearly a thousand skills on the platform. I’d be testing ordering a Domino’s Pizza using nothing but my voice (yeah, that took off…not!).

The problem was that I treated the opportunity like a set of instructions, not a problem space to think through.

I should have prepared like I was building something 0 to 1. Customer problem. Market context. Why now. What's the smallest thing we could build to test the idea.

Instead, I walked in with enthusiasm and not much else.

But I didn't stop there

Here's where this story takes a turn.

That pitch went badly. But I didn't let it kill the idea.

About a year later, in 2016, the Alexa Skills Kit was live and growing fast.

Developers everywhere were building skills — from trivia games to smart home controls.

And I thought, what if I built something useful internally?

So I did.

I wrote my first AI skill. It ran on Alexa and allowed people inside the organization to make inbound requests to me through voice.

Ten years ago.

A decade before everyone started talking about AI agents and automation, I was building a voice-powered intake system on Alexa.

It wasn't fancy. It wasn't going to change the world.

But it solved a problem—people needed a way to reach me, and I needed a way to manage inbound without drowning in emails and Chime messages (before Amazon grew up, admitted defeat, and used Slack).

The lesson for your career

I think about this story a lot when I'm coaching PMs.

Because the lesson isn't "always say yes" or "never walk into a room unprepared."

It's more nuanced than that.

The opportunities will come.

New technology. A new product area. A chance to pitch leadership. A stretch project that scares you.

The question isn't whether to say yes.

It's how you say yes.

Connect it to something real.

Don't pitch an idea in a vacuum. Tie it to a problem your audience cares about. If I'd walked into that room with Greg and said, "Here's a customer pain point, and here's how voice could solve it," the conversation would have gone very differently.

Prepare like it matters.

Even if you're early. Even if the idea is rough. Do the homework. Bring data. Bring a customer quote. Bring something that shows you've thought beyond the surface.

Don't let a bad pitch kill a good instinct.

My instinct about voice and AI was right. The timing and execution were off. But I came back to it. Most people would have just moved on.

That last one is the big one.

In a PM career, you're going to have meetings that don't go well. Pitches that fall flat. Ideas that get shut down.

The PMs who grow fastest aren't the ones who never fail.

They're the ones who fail, learn the lesson, and come back with something better.

So here's my question for you

Is there an opportunity sitting in front of you right now that you've been hesitant to grab?

Maybe it's a new technology you want to explore. Maybe it's a leadership conversation you've been avoiding. Maybe it's a role that feels like a stretch.

Don't wait until you feel ready. You won't.

But don't walk in empty-handed either.

Do the prep. Connect it to something real. And if it doesn't go perfectly... come back and try again.

That's how careers are built.

Rooting for you,

James