Insights
Why did the AI tool I bought stop being impressive?
Because the category moved, not because your tool broke. AI tools — and the expertise around them — are turning over on something like a 90-day cycle right now, so what you bought was state of the art the month you signed up, and a snapshot ever since.
That's a strange experience for an owner. Everything else you buy for the business — a truck, a compressor, a phone system — depreciates over years. This depreciates over a quarter, and it feels like you got taken. Mostly, you didn't. The tool does exactly what it did on day one. Everything around it got better.
That answer is the diagnosis. What follows is the treatment plan: how to tell disappointment from actual failure, how to audit the tool you're already paying for, and what to do with the verdict — because "it stopped being impressive" and "it stopped working" are different problems with different price tags.
What does "the category moved" actually mean?
Concretely, three things happened around your tool while it stood still. The engines improved — the underlying AI that powers nearly every tool in the category takes big steps a few times a year, so responses that felt startling at signup read as stiff next to what current engines produce. The features migrated — capabilities that justified your tool's price got absorbed as checkboxes into bigger platforms, which is one of the three gears that keep this whole category churning. And your own eye sharpened — the demo that amazed you in March can't amaze you in September, because you've spent six months watching AI do things. Your calibration moved. The tool didn't.
None of those three is a malfunction. Which is exactly why the right response isn't automatic replacement — it's an audit.
How do I audit the tool I already have?
One question, asked with the receipts open: is it still doing the job I bought it for? Not "does it still feel impressive" — feelings just got recalibrated, see above. Go back to why you signed up. There was a leak: missed calls, quiet estimates, slow invoices. Pull the last month and count what the tool handled — calls caught, texts answered, follow-ups fired, whatever its lane is. Three verdicts are possible.
The leak is still plugged. The unglamorous common case. The tool catches what it always caught; nothing about a category moving un-catches it. What changed is prestige, and prestige was never on the invoice — outcomes, not impressiveness, were the purchase. Keep the receipts from this audit, incidentally: a one-page count of what the tool handled last month is the exact artifact that makes next quarter's re-audit a ten-minute job instead of a fresh investigation.
The leak was never really plugged. The audit's uncomfortable second finding, more common than anyone advertises: the counting reveals the tool never handled much of what it promised, and the demo carried the sale. The fix isn't the newest competitor — it's going back to which leak you actually needed plugged, then buying against that with your eyes open. This is also the moment to reread the fine print on what leaves with you if you cancel, because that answer shapes your options more than any feature list.
The leak was plugged, and has started seeping. The genuine churn casualty: reply quality degrading against modern expectations, customers noticing the stiffness, competitors' response speed resetting the local bar. Real, but rarer than the feeling suggests — which is why you count before you conclude.
When is switching actually worth it?
First, a word about the pull you're feeling, because it's engineered. Every quarter's new tools demo against your current recalibrated eye — that's their whole job — while your existing tool demos against nothing; it just quietly works or doesn't. The comparison is rigged at the showroom door, which is precisely why the decision has to move from the showroom to the ledger.
Run it as arithmetic, not as mood, because switching has costs the shiny demo never itemizes: setup and reconnection time, retraining whoever touches it, the conversation history and configuration that may not survive the move if you were renting the workflow, and a fresh round of edge-case discovery on your customers' patience. Against that, price what the seepage actually costs you monthly — jobs lost to stiff replies, minutes lost to workarounds, using the same leak arithmetic that justified the tool in the first place. Seepage cost clearly exceeding switch cost: switch without sentiment. Otherwise: stay without embarrassment, and put the audit on a calendar to re-run next quarter. A tool you audited and kept is a decision. A tool you kept because auditing felt like work is a drift.
But outcomes don't churn — and that's the pivot. The tool stopped being impressive; the problem it was solving didn't stop existing. Calls still come in when nobody can answer. Follow-ups still don't send themselves. If your tool is still catching what you bought it to catch, it's doing its job, whether or not it demos well anymore. If it isn't, the fix is not automatically the newest replacement — it's deciding who should be responsible for keeping the underneath current. The owners who get this right stop buying tools and start assigning the churn: the quarterly turnover becomes a line item somebody else watches, instead of a surprise you keep discovering.
The audit above is yours to run, this week, with numbers you already have. The standing version — someone watching the underneath so the surprise never reaches you — is the work I do. If the count comes back ugly, or you'd just rather never run it again yourself, bring me the tool and the leak it was meant to plug: book a conversation.