Somewhere in your budget right now, there's a line item for AI. Maybe it's a Copilot rollout. Maybe it's an internal chatbot pilot. Maybe it's a vendor contract for "AI-powered" something that a department head championed a year ago.
And somewhere in your head, there's a quiet, uncomfortable question you haven't said out loud in a leadership meeting: did any of that actually work?
I get called in after this exact moment more often than at any other point. Not before the AI spend — after it, when the results didn't show up and nobody wants to be the one who admits it.
The confusion at the root of it
Here's the mistake, plainly: buying AI tools is not the same thing as having an AI strategy. It feels like progress. It shows up as a bullet point in a board update. It is not the same thing, and the gap between the two is where most AI budgets quietly go to die.
A tool is a capability. A strategy is a decision about which business outcome you're pointed at, who's accountable for it, what "working" looks like in numbers, and what you'll do differently once you have it. Most companies buy the capability and skip every other part of that sentence.
I've sat in enough of these post-mortems to see the pattern repeat almost exactly:
- Someone senior feels pressure to "do something on AI." Competitive anxiety, a board question, a conference talk — doesn't matter which.
- A tool or vendor gets selected, often based on a demo rather than a defined problem.
- A pilot launches in one team, with enthusiasm but no clear success metric.
- Six to twelve months pass. The pilot is still "a pilot." Usage is patchy. Nobody can say with a straight face what it changed.
- The initiative quietly stops getting mentioned in leadership meetings.
None of that is an AI problem. It's a strategy problem that happened to involve AI.
The question that exposes it every time
When I'm brought in to look at an existing AI initiative, I ask one question first, before anything technical: "What metric were you trying to move, and what's it done since you started?"
You'd be surprised how rarely there's a clean answer. Sometimes there's no metric at all — just a general sense that "efficiency" should improve. Sometimes there's a metric, but nobody's actually tracked it since launch. Sometimes the metric moved, but for reasons unrelated to the tool, and nobody separated the two.
If you can't answer that question cleanly for your own AI initiative right now, that's not a character flaw — it's the single most common gap I see, and it's completely fixable. But it does mean what you have today is activity, not strategy, no matter how much was spent getting there.
What separates a strategy from a tool rollout
A few things are almost always missing when a company has tools but not a strategy:
- A named business owner, not just an IT or innovation owner. Someone whose job depends on the outcome, not just the deployment.
- A single, specific metric the initiative is supposed to move — cost per ticket, cycle time, error rate, revenue per rep — something you were already measuring before AI showed up.
- A baseline, taken before the tool launched, so "improvement" means something other than a feeling.
- A decision point — a date and a threshold at which you decide to scale, adjust, or kill the initiative. Without this, pilots just live forever in pilot purgatory.
- A change to how work actually gets done, not just a tool sitting alongside the old process, used by whoever remembers to use it.
If your current AI effort is missing three or more of these, that's your diagnosis. It's not that AI doesn't work for a business like yours — it's that nobody built the scaffolding around it that turns a capability into a result.
The good news
This is one of the more fixable problems I encounter, precisely because the hard part — getting leadership willing to invest in AI at all — is already done. You don't need to convince anyone AI matters. You need someone to come in, look at what's actually there, and turn "we have some AI tools" into "we know exactly what this is supposed to do, for whom, and how we'll know if it worked."
That's usually a matter of weeks, not months, and it doesn't require ripping out what you've already bought. It requires putting a decision structure around it that should have been there from day one.
If that sounds like where you are — tools in place, results unclear, and a quiet suspicion that something's missing — let's talk it through. I'll tell you plainly whether what you have is salvageable and what I'd change first.