The worst advice a small business owner gets about AI is “keep up.”
You can’t. I say that as someone who works with these tools all day, every day. The model that impressed everyone in March looked ordinary by August. The tool somebody’s nephew swore by in spring was abandoned by fall. If your AI strategy is chasing what’s new, you have signed up for a treadmill that speeds up every quarter, and your actual business, the one with customers and payroll, gets your leftover attention.
So, I use a different filter, and it’s the opposite of keeping up. Before I point AI at anything, I ask one question: will this problem still exist in 10 years?
Not the technology. The problem. Because here’s the thing the hype cycle never says out loud: AI changes how problems get solved. It almost never changes whether they exist. Homes don’t stop breaking because a new model shipped. Companies still need to know who they’re really hiring. A missed call doesn’t stop being a missed customer. Those problems were here before AI, and they’ll be here after the current tools become museum pieces. That permanence is exactly what makes them worth building on.
I spent 30 years leading enterprise sales organizations. At 56, I run four businesses by myself from a home office in McKinney, Texas, and every one of them was chosen with that filter. Let me show you what it looks like in practice, because my businesses can argue better than my opinions can.
The first is a home repair service that runs entirely over text messages. A homeowner texts us what broke. From that one text, an AI agent reads the request, scopes the job, prices it, matches a vetted local contractor, works out scheduling with both sides, and handles payment and follow-up. It serves homeowners in two states today, with no dispatcher and no office.
Notice what the bet was. It wasn’t a bet on chatbots or any particular model. It was a bet that homes will keep breaking and that people will always take the easiest trusted path to getting them fixed. The AI changed how that problem gets solved. It did not change whether the problem exists. That’s the filter doing its work.
The second business helps companies figure out who they’re actually hiring. An AI conducts a structured interview and turns the conversation into evidence a hiring manager can check, instead of a resume nobody can. The durable problem underneath it is older than I am: every company that has ever existed has needed to know who it is really bringing aboard. Interviews used to depend on one manager’s gut on one afternoon. The gut was the changeable part. The need was not.
The third answers calls when a small business can’t. It responds to inbound leads in seconds, any hour, and books the appointment before the prospect moves down the list. The permanent fact underneath: a missed call has always been a missed customer, since the day businesses got phones. Nothing about smarter software makes that less true. It just moves the response time from tomorrow morning to right now.
Three businesses, one filter, restated plainly each time on purpose: find the problem that will never go away, then point the newest tools at it.
Now here’s your version, and I want to be honest about how little it asks of you. This filter does not require you to become a technologist, rebuild your operation, or bet the company. Most owners I talk to are too busy running the business to take on a transformation, and they shouldn’t have to. The move is smaller than that. You already know your durable problems. You’ve been living with them for years. They’re usually the most boring things in the shop.
Ask yourself three questions. What do my customers need from me that they needed 10 years ago and will need 10 years from now? What breaks or repeats every single week no matter what I do? And where does a slow response cost me money? Your answers are your list. The estimate that takes four days to send. The phone that rings into voicemail at 5:01. The follow-up that only happens when things are slow. The new hire you can’t evaluate until you’ve paid them for a month or two.
Then pick the most boring one and point AI at that. Not at content calendars, not at whatever launched this week. At the permanent problem. Boring is a feature here. Boring means the problem has survived every technology wave so far, which is the strongest evidence it will survive the next one, with you still getting paid to solve it.
There’s a hedge built into this that I think about a lot. A bet on novelty expires the day something newer ships, and something newer always ships. A bet on a permanent problem compounds. When the tools improve, and they will improve faster than any of us can track, a business aimed at a durable problem doesn’t become obsolete. It becomes more efficient at something people still need. The improvement lands in your pocket instead of making your strategy stale. You don’t have to outrun the technology. You just have to be standing on ground that doesn’t move.
The owners actually winning with AI right now are not the ones who know the most about AI. I know some of each kind. The winners are the ones who know exactly which of their problems will never go away, and had the nerve to hand those problems to the machine first. The hype will keep changing channels. Homes will keep breaking. Phones will keep ringing. Build on the part you can count on.
Wade Kerzie spent 30 years leading enterprise sales organizations and now runs four businesses solo from McKinney, Texas, using AI: a home-repair-by-text service, an AI hiring engine, a lead-response system for small businesses, and Kerzie AI Solutions, a consulting practice. He writes at kerzie.ai.
