Quick question. Have you had this week yet?
You read something about AI. You think okay, I really should get on this. You open one of the tools. Type something in. Get back something that reads like a robot wrote a LinkedIn post. Close the tab.
Then you get back to actually running your business.
And there’s this little hum in the background: I should be doing something about that.
That hum’s expensive. Not because you’re behind — you’re less behind than you think — but because most owners diagnose it as a willpower problem. It isn’t. It’s three things, and once you see them it stops being mysterious.
Number one: time
Your day is customers, staff, suppliers, and whatever caught fire this morning.
The gap between “I should look into AI” and actually having it working isn’t a Saturday. It’s setting it up, learning what good looks like, redoing the stuff that isn’t, and then keeping up as it changes every few weeks.
That’s a second job. And there’s no week where that fits — not because you’re disorganised, but because your time’s already spoken for by work that pays this month. Any plan that needs you to magic up ten hours a week isn’t really a plan.
Number two: technology
Say you found the time. You’d hit the next wall about an hour later.
Which tools? How do they talk to the stuff you already use? Why does one prompt give you something usable and another gives you waffle? How do you make it sound like you instead of a corporate newsletter? And how do you keep any of it working when everything changes again in six weeks?
This part is genuinely hard. Which is worth saying, because a lot of owners quietly feel a bit thick about it. You’re not missing something obvious. It’s a real skill, and it’s moving fast.
Number three: talent
Fine, you think. I’ll just hire someone.
Then you look at what that costs. In the US an AI manager averages around $103,000 a year. The ones who actually know what they’re doing tend to sit between $130,000 and $180,000 — before benefits, gear, software and recruiter fees.
For most small businesses that’s not a decision, it’s a non-starter. And even if the money worked, you’d be bidding against tech companies, interviewing for a skill you don’t have, and hoping they don’t leave in eighteen months with everything you taught them.
Now here’s the annoying bit
Look at what’s on offer to fix this.
- A course or challenge. Takes your time (problem one), assumes you’ll pick up the tech skill (problem two), and makes you the talent (problem three). So it fixes nothing and adds homework.
- An AI tool. Gives you tech. Still needs your time to run it and your skill to run it well. One problem solved, two made worse. It’s why so many AI subscriptions get quietly cancelled around month three.
- An agency. Some skill, some tech, plus a whole new thing to manage. Briefing, chasing, reviewing.
- A VA. Some hands — but you’re training them, and you’re supplying the tech. Which needs the time and skill you didn’t have to start with.
Spot the pattern? Every single one hands part of the job back to you.
And you’re the bottleneck. That’s why nothing’s stuck.
The question that changes it
Nearly everyone asks: how do I learn AI and get it going in my business?
There’s no good answer to that if you’re running a business.
Try this instead: who’s running AI in my business?
Now watch the three problems. Time — not yours. Tech — someone else’s job to keep up with. Talent — you’re not hiring and training anyone, you’re getting someone who’s already trained.
All three, gone.
One catch, and it matters: that only works if the answer includes the person and the tech and the plan. Miss one and it lands right back on your desk. Which, funnily enough, is exactly why the tool went unused and the VA needed training you didn’t have time to give.
So where does that leave you?
There’s a role forming in every serious business right now, and big companies are paying six figures for it. Small businesses need it just as much and can’t hire like that.
Which leaves a fairly obvious question — and it’s the right one to sit with. If you need the role but can’t justify the salary, and one person on their own wouldn’t cut it anyway, what’s the alternative?
Grab the free guide: The 7 Places AI Should Be Working In Your Business →