The Middle Third: Why Your AI Rollout Will Succeed or Fail Based on People You're Ignoring

I've got 20 minutes on stage at the owners meeting to talk about AI. And I keep catching myself writing the talk for the wrong people.

The room splits into thirds — and if you're rolling out AI (or honestly, any transformative technology) in your firm, yours does too.

One third already lives in this. Claude Code open on a second monitor right now, three steps ahead of my slides. I'm not teaching them anything in 20 minutes.

One third has decided. AI is hype, or a threat, or both. I'm not moving them in 20 minutes either. Twenty minutes against a made-up mind is just me losing on a schedule.

Then there's the middle third. Curious. A little nervous. They've heard "AI" 400 times this year and still don't know where to put their hands. They want in. They can't find the door.

That middle is the entire game.

The Rogers Curve Isn't a Prediction — It's a Warning

Everett Rogers published Diffusion of Innovations in 1962, mapping how new technology spreads through populations. The curve is famous: innovators, early adopters, early majority, late majority, laggards. We frame it as destiny — adoption flows left to right like water finding level.

But the curve doesn't move itself. Someone has to move it.

I learned this running security awareness programs for fifteen years, back when "don't click suspicious links" was a radical concept. You'd get the budget, build the training, launch the program — and then make a choice that determined everything that followed.

You could spend your time on the guy who clicks every phishing link out of spite. The one who forwards IT security emails to his personal Gmail "in case he needs them later." The one who tapes his password to his monitor because "nobody explained why this matters."

I tried. God knows I tried. You know what converting a committed skeptic gets you? Nothing. Maybe you move one person. Maybe. The curve doesn't budge.

Or you could spend that same budget on the 60% in the middle who actually want to do the right thing and just don't know what it is. The people already leaning who just need someone to point.

Spend your time there and the whole curve moves.

What the Middle Third Actually Needs (And Why You're Probably Not Giving It to Them)

Here's what I cut from my AI talk in the last revision:

The demo that makes power users nod knowingly. Gone.

The debate slides about AI limitations and failure modes that skeptics would've loved to fight about for 45 minutes. (My favorite slide. A beautiful breakdown of where ChatGPT hallucinates in tax research contexts. Still stings.) Gone.

The architectural diagrams. The cost-benefit frameworks. The "here's our 18-month roadmap" timeline. All gone.

What's left is one thing: here's the first useful thing you can do Monday, and here's exactly where to click.

Not "AI can transform your workflow." Not "imagine a future where." Not even "here are five ways partners are using this."

Click here. Type this. Hit enter. Here's what good looks like.

The middle third doesn't need inspiration. They're already curious — that's why they're in the middle. They don't need to be convinced AI matters. They've heard that 400 times this year.

They need the door. They need to know where to put their hands.

Nobody Gets Fired the Day the Technology Arrives

I watched this pattern play out with cloud adoption in professional services. Remember when "our data can't leave our building" was the default position?

The firms that moved fastest didn't do it by converting the security partners who'd built their reputation on physical control. They didn't do it by impressing the CTOs who were already running shadow AWS instances.

They did it by making it so stupidly easy for the middle — the project managers, the senior associates, the people doing actual client work — that the path of least resistance became the new way.

Box. Dropbox. Tools so simple that the "here's how" conversation took 90 seconds. The middle third moved. Then the curve moved. Then the skeptics looked around and realized they were the only ones still burning CDs.

Nobody gets fired the day the railroad arrives. The town just slowly empties out. But somebody has to lay the first hundred feet of track where people can actually see it, touch it, use it.

The Mistake I Keep Watching Firms Make

They design the rollout for the extremes.

Big all-hands presentation. Impressive demos. The head of innovation (if you have one of those) shows what's possible. ChatGPT writes a memo. Claude analyzes a contract. Copilot generates working Python.

The power users are bored — they've been doing this for months.

The skeptics cross their arms — see, it's just a party trick, it got the case cite wrong, this proves my point.

The middle third leaves the room thinking "that's cool" and having absolutely no idea what to do Tuesday morning. The launch changes nothing because it gave the middle inspiration instead of ignition.

I did this wrong for years in security. Beautiful training programs. Engaging videos. Gamification, even — points for spotting phishing attempts, leaderboards, the works.

Know what moved the needle? A one-page PDF that showed exactly which button to click when you get a suspicious email. Not why. Not the theory. Not the risk framework. Just: see this? Click here. Done.

The guy who was already forwarding weird emails to IT didn't need convincing. He needed the path.

The Question That Matters for Your Monday Morning

Here's what I'm not saying: ignore the power users and abandon the skeptics.

The early adopters are your proof of concept. Let them run. Get out of their way. Capture what they learn and hand it to the middle.

The skeptics keep you honest. They ask the questions about accuracy, liability, client confidentiality that prevent you from rolling out something half-baked that blows up in your face. Listen to them. Just don't spend your rollout budget trying to move them.

Your success lives in the middle third. The ones who are curious. A little nervous. Want in. Can't find the door.

So here's what to ask yourself — or your innovation team, or whoever's driving your AI rollout, or honestly any significant technology change in your firm:

Who's in that middle right now? Name three people.

And what's the one stupidly simple, Monday-morning thing you're putting in front of them? Not a vision. Not a roadmap. Not a 40-minute training.

The one thing they can do in five minutes that makes them say "oh, THAT's what this is for."

Because you don't move a firm by converting deniers or impressing believers.

You move the people in the middle who are already leaning.

Point them at the door. Then get out of the way and watch what happens.

What to do Monday morning: Identify your middle third. Pick ONE task AI can make measurably easier for them — not impressive, easier. Write the literal three-sentence instruction: click here, type this, here's what good looks like. Send it. See who uses it. That's your real adoption curve.