Ask a room of business owners whether their company uses AI, and nearly every hand goes up.
Ask them what it has actually returned — in hours, in margin, in work that gets done that didn’t before — and the room goes quiet.
The National Bureau of Economic Research put that question to nearly 6,000 senior executives at the start of this year. Seventy percent of their companies actively use AI. Nine in ten reported no measurable impact on productivity or employment over the previous three years.
That gap isn’t a rounding error. It’s the entire story of this market right now. And if you’re being honest with yourself, you probably already know which side of it you’re on.
Activity Is Not Maturity
Here’s the trap. Usage is easy to see and easy to feel good about. People are using the tools. Someone drafted a proposal faster. Someone summarized a call. The dashboard, if you had one, would look busy.
But busy isn’t the same as better. What you have in most companies is what we’d call random acts of AI — individual people, solving individual problems, in whatever tool they landed on, with no shared thinking behind any of it. It produces exactly what you’d expect: a lot of activity and no compounding.
And there’s a reason it stalls there. It isn’t budget. It isn’t that you bought the wrong tool. It’s that nobody handed you a picture of what good looks like, so there’s no way to know where you stand or what comes next. You’re hearing “agents” and “autonomous” and “cowork,” and none of it maps to anything in your actual business. Without a model, you can’t sequence. Without a sequence, you take steps that don’t build on each other.
That’s not a discipline problem. It’s a missing map.
What the Tenth Company Did Differently
The companies getting real returns didn’t find a better model or write better prompts. They did something much less glamorous: they got the order right.
AI value in an SMB rests on five things, and they don’t carry equal weight.
- Productivity opportunity. How much of your team’s week is going to work a machine could do — searching for a document that exists somewhere, retyping data between systems, writing the same status update for the eleventh time. This is where the hours are. It’s also the pillar most companies never measure, which is why they can’t tell you what AI has actually returned.
- AI in motion. How many people are actually using the tools — and what happens to a good idea when someone finds one. Does it travel, or does it die at that person’s desk? A perfect strategy in a company that doesn’t adopt is worth nothing.
- AI foundation. Can an AI assistant actually get to your data? This is the quiet one. If your active work lives on a local drive or an old file server, AI cannot help you with it — and it doesn’t matter which model you license. AI can only work with what it can reach.
- Vision and leadership. Someone has to name a specific business outcome AI is supposed to move, and own it by name. “We should be doing AI” is not an outcome.
- Trust foundation. The right people seeing the right information, and a team that knows what’s okay to put into a tool. Not a governance project — a floor.
The companies seeing returns aren’t stronger across all five. Nobody is. They just knew which one was holding them back and fixed that one first.
Your Weakest Pillar Is Not Your Failure
This is the part that changes how the whole thing feels.
When companies discover their AI foundation is weak — files scattered, half the team on the wrong license, collaboration split across three different tools — the instinct is embarrassment. It reads like a report card, and the reaction is to go quiet and hope nobody looks too closely.
That’s the wrong read entirely. Your weakest pillar isn’t the thing you got wrong. It’s the thing with the most upside left in it. It’s the highest-leverage move on the board, and it’s usually the one you can act on fastest, because the gap is obvious once you can see it.
The shape of your profile matters far more than the score. Two companies can both sit at 40 out of 100 and need completely different next steps — one needs licenses and file migration, the other needs a reason to care. Averages hide that. The imbalance is where the value is.
In 34 years working alongside Utah business owners, we’ve watched a lot of tools come and go. The businesses that get the most out of any new technology aren’t the ones with the biggest budget. They’re the ones who knew exactly where their weak point was before they spent a dollar fixing it.
The Step in Front of the Step
Everyone is being told to do AI. Almost nobody is being told what to do first — which is why so much of the spend evaporates.
You don’t need a strategy deck. You don’t need a governance framework. You don’t need an answer for agents. You need an honest read on where you stand across those five pillars, and one reachable first step — the kind you can start on Monday and feel by the end of the month.
That’s it. That’s the whole difference between the nine and the one.
Most companies in this position are sitting on somewhere between four and eight hours per knowledge worker, per week, locked up in repetitive admin and information search. Not theoretical hours. Hours your people are spending right now, this week, on work that doesn’t require them.
The question isn’t whether that time exists. It’s whether you know where it’s hiding.
Find Out Where You Actually Stand
The AI Readiness Assessment is a 12-question, three-minute diagnostic. You’ll get a score out of 100, a breakdown across all five pillars, and a prioritized 90-day plan built around your highest-leverage gap — not a list of everything that’s wrong.
No login. No sales call required. The report is yours either way.
[Get Your Score & 90-Day Plan →](https://lp.itnow.net/ai-readiness-assessment)
