Guides
Both are happening at once. Software you already pay for is gaining AI features you did not ask for, sometimes with a higher renewal attached. Separately, a small number of specific jobs in a small business can now be handled well by AI that was chosen for that job. The first kind arrives whether you want it or not. The second kind only works if someone decides what it is for.
We are a company with AI in its name, so read the rest with that in mind. What follows is the distinction we would want an owner to have before anyone sells them anything, including us.
Ask an owner what AI has done for their business and the honest answer is usually that it showed up uninvited. A feature appeared inside the scheduling platform or the job software. It was announced as an improvement. Sometimes the renewal went up in the same season. Nobody asked for it, it does not touch the thing that actually costs the business its Thursday afternoons, and it joins the reminders and the waitlist and the intake forms in the long list of paid features nobody has switched on.
Owners talk about this in public constantly, and the tone is not confusion. It is irritation at being charged for a direction they did not choose. If that is your reaction too, it is a reasonable reading of the evidence in front of you, and it is worth keeping.
We are not going to argue you out of it. The suspicion is doing its job. What it cannot do on its own is tell you which of the two things you are looking at.
An AI feature built into a platform was built for every customer that platform has. It has to serve the thousand person company and your four room practice from the same code, which means it answers a question the vendor decided was common. Whether that question is the one costing you your evenings is a coincidence. It might be. Most of the time it is not.
The other kind starts from the other end. You name the thing that eats the week. Someone then works out whether AI can do that specific thing reliably in your business, what it would cost, what it would still get wrong, and who carries it when it does. Sometimes the answer is that it can and it is worth doing. Often the answer is that a setting in software you already pay for does the job better, and the honest recommendation is to change the setting and spend nothing.
Those two things share a word and almost nothing else.
Including ours. These are the questions we ask before recommending anything, and they are quick enough to use in a sales call.
If the answer is a capability rather than a job, it has not been thought about yet. "Summarizes your data" is a capability. "Tells you on Monday morning which customers did not rebook" is a job. Only one of those can be checked against your calendar.
Everything in this category has a failure mode. Anyone who will not name theirs either does not know it or is not saying. The failure mode matters more than the feature, because when it happens it is your customer on the other end and your name on the message.
Something you can switch off costs nothing to try. Something that reorganizes your records, or that your renewal price now depends on, is a decision rather than a trial. Ask which one it is before you agree, not after.
Short list, because the honest list is short. Each one carries its limit, since the limit is where the bad advice hides.
Someone who writes at 9pm gets a real answer in your voice, with the next open slot. The limit: it books and answers; it does not diagnose, quote, or promise on your behalf.
It knows who has drifted, when that usually happens in your business, and sends the note you would have sent if you had the hour. The limit: you approve the wording once; it does not invent offers.
A slot that opens goes to the waitlist in minutes instead of sitting empty. The limit: a waitlist has to exist, which is often the first thing that gets set up.
Forms go out ahead of the visit, come back complete, and land in the record. The limit: the record has to be clean first, and after a rushed migration it usually is not.
Estimate follow ups, post visit notes, review requests, drafted and sent without anyone remembering to do it. The limit: it sends what was approved; it does not freelance.
A short note on how the week went: what came in, who did not rebook, where it leaked. The limit: it reports; the decision is still yours.
There is an assumption that this only matters to someone working alone. The opposite is closer to true. A solo practitioner loses time to admin, and it is visible to them every day. A practice with five practitioners and a front desk, or a field business running several crews, loses it in the handoffs: the same customer details typed into two systems, the job that falls between two people, the schedule that lives in one person's head and stops when they take a week off.
Those businesses can also put a number on it, which the solo operator often cannot. When an owner counts the hours their staff spend on documentation, chasing claims, and re-entering data, the figure is regularly a third of a payroll producing nothing. That is a different conversation from an annoyance.
The test is not headcount. It is whether anyone on the payroll is a systems person. If nobody's whole job is making the software serve the business, the software drifts, and it drifts faster the more people depend on it.
We look at your business, on its own data, and write down the AI work worth doing there, and what to leave alone. That document is free, it carries every cost including ours, and it is yours whether you work with us or not.
If the answer is that nothing here is worth doing yet, the plan says that plainly. That has happened, and it will happen again. An owner who reads it and changes one setting instead of buying anything has still had the better outcome.
Usually not on its own. The test that settles it: can you say the job it does as one sentence about your week? Six things are working reliably in small businesses right now: answering after hours, bringing quiet customers back, filling cancellations, doing intake before the visit, writing follow up, and a weekly read of how the business did. A feature doing one of those, at a price you would have paid anyway, earns its place. A feature doing none of them is a renewal increase with a label on it. Which of the six is worth doing in your business depends on where your time actually goes, and that is what the free plan works out.
Four, and they are worth pricing side by side before you react. Stay and absorb it. Stay and negotiate, which works more often than owners expect at renewal. Move to a competitor, which carries a migration cost that is usually underestimated. Or stay and switch on the features you already pay for, which costs nothing and is the option most owners have not tried. A price rise is a good moment to find out what the platform can already do.
No. The pain grows with headcount, because the work falls between people. What matters is whether anyone on the payroll is a systems person. A business with its own operations or IT function already has this coverage, whatever its size. A business without one does not, however many people are on the schedule.
Usually not, and we will tell you when the answer is no. Leaving a system that works to get an AI feature is one of the clearest bad trades in this category. The migration costs real money and real weeks, and the feature is rarely worth either. When a move is genuinely the right call, it is because the current system cannot do something the business needs, and the plan says so with the reasoning shown.
