Your Team Isn’t Using AI. Here’s Why That’s Your Fault.

A three-phase framework for building an AI-first culture — starting with the uncomfortable realization that you’re the only one who’s been experimenting

Tuesday morning. Team standup. I’m describing a skill I built that pulls meeting context from three different systems, assembles a briefing doc, and drops it into my notes before every client call. It saves me thirty minutes of prep, five times a week.

Silence.

Not the “that’s impressive” silence. The “I have no idea what you’re talking about” silence. The kind where people smile and nod because asking a follow-up question would reveal how far behind they are.

I’ve been building automations with AI tools for months. Skills that handle meeting prep, research workflows, content pipelines — the kind of work that used to eat my mornings. I assumed the team was doing the same. Different tools, maybe. Different use cases. But experimenting.

They weren’t. Not even close.

The gap nobody talks about

Here’s what I’ve learned from running a startup studio and advising early-stage founders: the gap between AI-forward leadership and AI-absent teams is widening faster than anyone admits.

Founders and operators are deep in these tools. They’ve felt the shift — the moment you realize that a task that took two hours now takes fifteen minutes, and the quality is better. That moment rewires how you think about work.

But the team hasn’t had that moment. They’ve seen the demos. Read the blog posts. Heard the hype. And they’re still doing everything the same way they did eighteen months ago.

This isn’t a training problem. You can’t solve it with a lunch-and-learn or a company-wide Slack message about the latest ChatGPT feature. It’s a muscle memory problem. And the only way to build muscle memory is to use the muscles.

Why mandating tools doesn’t work

My first instinct was wrong. I almost said: here are the tools, here are the workflows, go build these specific automations.

That would’ve been a mistake.

When you mandate a tool, you get compliance. When you create the conditions for experimentation, you get curiosity. Compliance produces people who can follow your automation playbook. Curiosity produces people who see their own workflows differently — people who start asking “why am I still doing this manually?” without being prompted.

The difference matters because the goal isn’t to get your team to use AI. The goal is to get your team to think differently about what’s repeatable, what’s automatable, and where human judgment actually matters. That shift in thinking is worth more than any individual skill or workflow.

Three phases, six months

What I’ve landed on is a phased approach that respects how people actually learn new capabilities. Not a training program. A culture shift with structure.

Phase Zero is about curiosity.

Everyone picks a project from their actual work — not a side hobby, not “plan my vacation with ChatGPT” — and uses whatever AI tool they want to make it better, faster, or automated. Any tool. Any workflow. The only constraint is that it has to connect to real work they do, and they have to show it to the team in thirty to forty-five days.

The rules are deliberately loose. Three to five hours a week. Use whatever tool interests you. Build for yourself, not for the company.

But there’s one structural piece that makes this work: a shared Slack channel where everyone drops a Friday update. What they tried. What broke. What surprised them. This does two things. It keeps momentum going so nobody procrastinates for a month and scrambles at the end. And it normalizes failure — because half of what people try won’t work, and the team needs to hear that out loud.

The show-and-tell at the end of Phase Zero isn’t a demo day. It’s a knowledge transfer. One person’s discovery becomes everyone’s shortcut. The engineer who figured out how to automate code reviews saves the whole team from learning that lesson from scratch. The ops person who tried to automate client onboarding and discovered the workflow was too messy to automate — that’s equally valuable, because now the team knows where the boundaries are.

Phase One is about observation.

Once people have the intuition from Phase Zero — once they’ve felt the tools, hit some walls, and understand what’s possible — Phase One asks a harder question: what in your actual workflow is repeatable enough to automate?

The first step is deceptively simple. Write down your repeatable tasks. Not vague (”do reporting”) and not granular (”copy this cell to that sheet”). For each one: What triggers it? What are the inputs? What are the steps? What’s the output? How often? What are the edge cases?

That last question is where most people stop too early. Edge cases are where automations break. And this is the insight I almost missed myself: the planning is more important than the building.

I’d put it at 75/25. Seventy-five percent of the work is understanding the workflow well enough to automate it. Twenty-five percent is the actual implementation. The natural instinct is to reverse that — spend a little time planning, then iterate. But if you skip the planning, you build something that handles the happy path and breaks everywhere else. You don’t think about what happens when the input is missing, when the data is messy, when the exception is actually more common than you thought.

The AI tools are good enough now that the building part is often the easy part. The hard part is knowing what to build.

Phase Two is about systems.

Not every Phase One automation should graduate to Phase Two. Some are personal productivity wins that stay personal — and that’s fine.

But the ones that go horizontal — the ones that multiple people could use, that run frequently, that save meaningful time per use — those need to become infrastructure. That means decoupling them from one person’s machine, centralizing the data layer, making them configurable, probably giving them a proper interface beyond a chatbot.

Phase Two is where the engineering investment happens. Phases Zero and One are essentially free — people using existing tools on their own time. Phase Two is where you start making build-versus-buy decisions and allocating real resources. That’s a fundamentally different conversation, and you’ve earned the right to have it because you’ve already validated the use case.

The chief of staff example

Here’s one that’s already made the journey from Phase One toward Phase Two in my own workflow.

Before every client meeting, I have a skill that pulls context from our project management tool, checks the client’s goals, looks at my task list, and assembles a briefing doc. It tells me what’s outstanding, what’s relevant to this specific conversation, and what I should be ready to discuss. It lands in my notes before I walk into the room.

This started as a personal pain point. I was showing up to meetings underprepared — not because I didn’t care, but because the prep was scattered across four systems and I was running between calls. The automation solved my problem.

But here’s what makes it a Phase Two candidate: it’s horizontal. Our mentors could use a version of this. Any team member running client meetings could use it. The workflow is the same — pull context, synthesize, brief — even though the data sources and outputs might differ.

That’s the Phase One to Phase Two bridge. A personal hack that solves a universal problem.

The thing leaders get wrong

The biggest mistake I see — the one I almost made — is treating AI adoption as a technology initiative. Selecting tools. Running training sessions. Building an “AI strategy” document that nobody reads.

It’s not a technology problem. It’s a comfort problem. Your team is smart enough to use these tools. They’re just not comfortable enough to start.

Phase Zero fixes that. Not by teaching them AI, but by giving them permission to tinker, a reason to finish, and a safe place to share what they learned — including what didn’t work.

The leaders who get this right do one more thing: they participate. Not as sponsors. As practitioners. If you’re asking your team to experiment for five hours a week but you’re not showing your own work at the show-and-tell, it reads as a mandate, not a culture shift. Your willingness to show a messy, half-working automation signals something that no memo or strategy document can: this is how we work now.

The question to sit with

Here’s the honest version of what I realized on that Tuesday morning call: the gap between me and my team wasn’t about tools or talent. It was about reps. I had hundreds of hours of experimentation behind me. They had zero.

No amount of strategy fixes that. Only doing the work does.

So the question for any founder or operator reading this: When was the last time your team — not you, your team — built something with AI that surprised them?

If you can’t answer that, you don’t have an AI strategy problem. You have a Phase Zero problem. And the fix is simpler than you think: give them a project, give them a deadline, give them a Slack channel, and get out of the way.

Then show up at the end and show yours too.
_______

Interested in taking your startup to the next level? Wildfire Labs is looking for innovative founders like you! Don't miss out on the opportunity to accelerate your business with expert mentorship and resources. Apply now at Wildfire Labs Accelerator https://wildfirelabs.io/apply and ignite your startup's potential. We can't wait to see what you'll achieve!