I brought up AI in a team meeting a while back, expecting a conversation. What I got was a room of blank stares.
These are smart, capable people. Most of them just had no idea where to start — not because they couldn’t, but because they’d never had a reason to stop and think about it. They were heads-down in the work, and “go figure out AI” was one more thing on a list that was already too long.
That meeting changed how I think about the so-called AI skills gap. We talk about it like it’s a training problem, as if the fix is a course, a license, and a stern email from leadership. It isn’t. The gap on your team is a leadership gap first. And the first skills worth examining are your own.
TL;DR
- The AI skills gap starts with the leader, not the team. Before you ask “why isn’t my team using this?”, ask “what are my AI skills?”
- You don’t need to be an expert. You need to be curious, aware, and willing to go first.
- Read your team’s mindset before you push. Some are already building things you don’t know about. Others are skeptical, intimidated, or just tired of the hype.
- Real adoption comes from within. Let people solve their own pain points and share what they find. Forced adoption doesn’t stick.
- Move fast and responsibly. Put simple guardrails in place so people can experiment without fear, and so you can show you’ve moved quickly and carefully.
- If you do only one thing: pick a single task you still do the slow way and run it through AI this week.
The first gap is yours
Before you go looking for the gap on your team, look for it in the mirror.
When a leader says “we need more AI,” what do they usually mean? Too often, not much. It’s a directive without a foundation — an order to chase something the person giving the order hasn’t tried to understand. Do you actually know what these tools can and can’t do? The trade-offs of using them, and of not using them? How fast the ground is shifting under all of it? Are you genuinely curious about any of that, or are you just barking?
I don’t consider myself an AI expert. I’m not technical by trade. But I got curious, and curiosity turned out to be most of the job. I signed up for my own Claude account. I went from the free tier to Pro to Max inside a few weeks, because I was generating ideas and projects faster than I could keep them organized. That’s the real bar for a leader — not expertise, but a genuine willingness to be aware, mindful, curious, and adaptive. You can’t lead a team somewhere you’ve never been willing to go yourself.
Read the room before you push
Once you’ve done your own homework, the next move still isn’t a mandate. It’s a read of the room.
Your team is not one audience. They’re a spectrum. On one end are the people already moving faster than you realize. After that blank-stares meeting, I found out a contractor on my team had quietly been using AI to build a set of internal reporting tools, from scratch, and close to delivering. I didn’t assign it. I didn’t even know about it. That’s the eager end: people who never waited for permission.
On the other end are the folks who aren’t eager at all. Maybe they’re skeptical. Maybe the pace genuinely worries them. Maybe they’re intimidated, or just plain sick of hearing the word “AI” in every meeting. Often it’s some mix of all of it.
Your job isn’t to drag everyone to the same place at the same speed. It’s to meet the adoption curve where it actually is.
A team that's forced to do something won't do it well.
Stop hunting for a problem to fit the tool
This is where a lot of organizations go sideways. AI becomes the shiny object, and suddenly everyone is looking for a place to put it.
As a MarTech person, that instinct makes me wince. I’ve spent a career watching people fall in love with a platform and then go searching for a problem it can solve. AI deserves better than that, and so does your team.
The adoption that actually lasts, the kind that moves both productivity and engagement, comes from the team, not from a slide you presented. They have to generate most of the use cases themselves, because they’re the ones who know where their own work hurts. So start small and personal. Have people fix their own pain points first. Then let them start working together.
Here’s one I’m about to start: “Demo Fridays.” The idea is simple. A few times a month, we get together for a quick show-and-tell of whatever we’ve been using AI to do. It doesn’t have to be big. It doesn’t have to be code, or a tool someone built. The point is to make the wins visible — to spark someone else’s idea, and maybe kick off a collaboration that wouldn’t have happened on its own. Adoption spreads when people are excited to try, not told to.
What leading actually looks like
So beyond getting out of the way, what does a leader actively do? A few things.
Go first. I keep returning to this one because it’s the move that worked. I paid for my own account out of pocket and started quietly delivering reports and other work faster than my leaders expected. Then I showed them a tool I’d built on that same personal account. It was a calculated leap, not a reckless one, and once they saw real value, the appetite for broader use opened up on its own.
Leading with proof beats leading with a pitch.
There’s a smaller version of this you can run in a one-on-one. In a quarterly check-in, I once walked a team member through how I was actually using AI day to day — the workflow, the outputs, the rough edges. You could watch it land. They started rethinking their own routine and came back with an idea to improve part of our website experience. Their eyes lit up the moment the use case was theirs instead of mine.
Create slack space. This is the one I’d underline. People can’t learn a new way of working when every hour is already spoken for. If your team is putting in 40-plus hours just to stay even, “go experiment with AI” is a tax, not a gift. So I’ve started framing the goal plainly: I want these tools to buy us a little slack in the week — time to noodle on our real problems, play with ideas, and get to know how this stuff actually works. Shorten a few cycles, and you free up room for the innovation everyone keeps asking for.
Inspire instead of mandating. For the skeptics, I’m betting on gravity, not force. My team runs the gamut: deeply technical developers, designers and UX folks who think like artists, project managers. You don’t push that group. You let them get pulled along by people they respect. I’ll be honest about the other side of that coin, though: if someone keeps resisting long after the rest of the team has moved, it may be telling you something about long-term fit. That isn’t a threat. It’s just true.
Move fast — and responsibly
None of this works if it scares the people whose job is to keep the organization safe.
AI can feel like the Wild West, and that has to be unsettling for IT and security leaders. Threat vectors are multiplying from inside and outside the org, and “let everyone experiment” sounds, to the wrong ears, exactly like how data walks out the door. The answer isn’t to slow down. It’s to put up guardrails early, simple ones your team can actually remember, so people can move without fear and you can show your work later. At some point, being able to demonstrate that you moved quickly and responsibly stops being a nice-to-have. It becomes the price of admission.
If you lead a marketing team in healthcare, a starter set might look like this:
- Never put PHI or anything patient-identifiable into a public AI tool. This should already be muscle memory, but say it out loud anyway. Most marketing folks rarely touch that data — confirm who actually does.
- Treat confidential and competitive material the same way. Strategy, unreleased work, contracts: anything you wouldn’t email to a competitor stays out of consumer tools unless it’s an approved, enterprise account.
- Know the difference between consumer and business tiers. They handle your data differently. Use approved accounts for work, and understand what each one does and doesn’t do with what you type into it.
- Keep a human on the hook for the output. AI is confidently wrong sometimes. Someone owns the final product, especially anything public-facing or regulated.
- Bring IT and security in early, not after. You want them as partners who helped shape the rules, not a team that finds out later.
You don’t have to publish all of this to the whole company on day one. But write it down now. You’ll be glad it exists when someone asks.
One more expectation to manage while you’re at it: the leaders above you may want a hockey-stick jump in productivity the moment your team touches AI. It rarely works that way. Early on, people are learning, and learning costs time before it saves any. Set that expectation honestly, or the ramp-up will look like failure when it’s really just the cost of doing it right.
Where to start Monday
If you take one thing from this, take this: AI starts with you. You don’t have to love it. You do have to acknowledge it.
Not everyone will take the same leap I did, and that’s fine. But I’d bet someone on your team already has. Part of the job is just finding the innovation that’s already happening under your nose and giving it some air.
And if you want the smallest possible place to begin, pick one task you still do the slow way and run it through AI this week. Still writing strategy briefs from a blank page — or worse, “open up, save as” on last year’s version? Make AI your brief-writing partner. Still building decks one slide at a time? Let it help you shape the message and the flow. You don’t have to let it build the whole thing, though you can. Even a little help with the content and the wording can change how fast you move.
The gap was never really about skills. It was about who was willing to go first.