The Uncomfortable Gap Between AI Efficiency and Actual Growth
Here’s something I keep running into with the business owners and senior leaders I work with: their teams are using AI. Genuinely using it not just dabbling. They’re clearing backlogs, tightening operations, replacing tools, building internal workflows that would have taken months before. By every visible measure, things are moving.
And the top line hasn’t budged.
In my latest conversation with Matt Graham, he put it as cleanly as I’ve heard it said: everyone is using AI, but no one is making more money. He was being a little facetious businesses generally are becoming more profitable as costs come down. But the growth? The revenue expansion? The aggressive top-line results? Those aren’t showing up at the same rate as the efficiency gains.
The reason, when you pull it apart, is almost never about the technology.
You’re Optimizing the Wrong Things
Every business has two or three places where focused energy produces results that are completely disproportionate to the effort. Call them leverage points, bottlenecks, whatever language works for you the concept is the same. These are the constraints that, if you removed them, would actually change the trajectory of the business. Not make it smoother. Change it.
AI is not going to those places.
What AI is going to, in most organizations right now, is the backlog. The wish list. The pile of things that felt too hard, too time-consuming, or too low-priority to tackle before AI made them fast and easy. And clearing that pile feels like real progress because it is real progress. The podcast gets processed faster. The scheduling tool gets replaced. The internal documentation gets written. The data gets cleaned. Real work, real results.
But none of it is moving the business.
I talked with a wealth advisory founder this week who illustrated this exactly. He was frustrated because every AI solution being pitched to him was about operational execution making his back-office more efficient, automating compliance workflows, speeding up reporting. His response to all of it: I don’t have an execution problem. I don’t have enough clients to execute for. What I need is sales and marketing.
He needed AI pointed at customer acquisition. He was getting AI pointed at his backlog. Those are completely different problems, and no amount of operational efficiency solves a revenue constraint.
As Matt put it: it’s a fool’s errand to stabilize operations for scale when you’re not actually at scale.
Why Curious People Are Especially Vulnerable to This
This pattern is not random. It shows up most predictably in organizations led by curious, tech-savvy people and that includes most of the executives and founders I work with. When AI drops a new capability, the instinct of a curious leader is to go explore it. That instinct is one of their genuine strengths. It’s also exactly what makes this trap so easy to fall into.
Matt runs development teams and he sees it constantly: his engineers want to replace every third-party tool with something they built themselves. Before AI, this was a normal impulse he could manage. Now, with AI making it fast and cheap to build internal tools, the pull is much stronger. He’s had friends who ran dev shops brag about how efficient their internal tooling was while they had twelve clients and needed forty.
I’m not immune to this. On the morning we recorded this episode, Riverside the platform we use to produce the show dropped an MCP connector. My immediate instinct was to drop everything and spend the day playing with it. I knew, in real time, that this was a side quest. I said it out loud. And I still felt the pull.
The honest takeaway here is that curiosity is an asset that requires management. It produces genuine value when it’s directed. When it’s undirected, it fills available time with interesting work that doesn’t move the business.
The question that actually matters is this: if you mapped where your AI energy is going against the two or three constraints that are genuinely limiting your growth right now, how much of that energy would line up?
For most organizations, the answer is: not much.
What to Do Instead
The fix is not complicated, but it requires a conversation most leadership teams aren’t having.
Name the two or three highest-leverage points in your business right now. Not ten. Not a list. Two or three. The places where, if something changed, the business would look meaningfully different in six months. For some businesses that’s customer acquisition. For others it’s retention, or delivery capacity, or a single sales channel that’s underperforming. Every business is different, and no framework substitutes for the honest answer specific to your situation.
Then look at where your team’s AI energy is actually going. Not where it’s supposed to be going where it’s actually going, day to day, in the tools people are using and the projects that are getting time. Compare those two lists.
If they don’t match and in most organizations, they won’t that’s your strategy conversation. Not which AI tools to buy, not which workflows to automate next. The conversation about alignment between your AI resources and your actual growth constraints.
This is also a management problem in disguise. Curious team members will always find productive-feeling AI work to do. That’s not a character flaw; it’s how curious people work. The organizational leader’s job is to point that curiosity at the right targets, give it a sandbox with defined boundaries, and make the highest-leverage work feel as engaging as the interesting side work. That last part is harder than it sounds.
One thing worth naming directly: there is a real flight risk in organizations that get this wrong in either direction. Shut down experimentation entirely especially in highly regulated industries where leadership has decided AI is too risky to touch and you will lose your best, most curious talent to organizations that give them room to build. That talent is not coming back. The answer is not to remove guardrails; it’s to build guardrails that protect the three things that matter most while giving people real latitude everywhere else.
AI bought you time and capacity. The question is whether you’re spending it on the right problem.
If you want help figuring out where AI should actually be pointed in your business, I built a free resource for exactly that: https://assessment.ascendlabs.ai/
If you’d rather talk through your specific situation: grab time here. No pitch. Just a real conversation about where you actually are.
