How to tell which local businesses will actually buy AI

The best predictor that a local business will buy an AI system is not revenue, not how dated their website looks, and not whether the owner has heard of ChatGPT. It is payroll. The JPMorganChase Institute tracked AI payments across 4.6 million small businesses and found that by December 2025, 26.1 percent of employer firms had paid for an AI service against 15.3 percent of nonemployers. Employer firms under $250,000 in revenue adopted at 27.6 percent, which beat nonemployer firms above $250,000 at 19 percent.

So a three-van plumbing outfit doing $200k is a better prospect than a solo consultant doing $400k. That one fact should reorder your call list. Here is the rest of what I check before dialing, the signals people misread as objections, and the businesses I now skip on purpose.

Why does having staff predict the sale better than revenue?

Because somebody has to own the system after you install it.

A solo operator buys software and then has to implement it between jobs, which in practice means never. A business with three employees has a person who already answers the phone, already touches the calendar, and already gets blamed when a lead goes cold. That person becomes the internal champion for your build whether they volunteer or not. The JPMorganChase authors reach the same conclusion from the data, writing that the pattern suggests "having access to additional human capital to implement and integrate AI may be more important than financial resources alone."

The gap is also widening, which matters if you are picking a lane for the next year. In January 2023 employer firms were at 9.6 percent and nonemployers at 4 percent, a spread of 5.6 points. By December 2025 the spread had grown to 10.8 points. Employers are pulling away.

Note what that report counts. A business is an adopter if it has ever paid for an AI service, drawn from Chase Business Banking transaction data. Free ChatGPT accounts do not show up. These numbers describe businesses willing to put AI on a card, which is the population you are hunting for.

The four things I check before I dial

None of this takes more than ninety seconds per business.

Do they have employees. The website usually tells you. A team page, a "careers" link, "our technicians", photos of more than one person in a branded shirt. If it is one name and one cell number, keep moving unless you have a reason.

Does a missed call cost them a job. Call at 6:40pm and listen to what happens. A voicemail box that is full, a ring-no-answer, or a generic answering service reading a script are all money leaving the building. A business where the phone is the till will follow your pitch without translation. A business where leads arrive by email referral will not.

Is the revenue appointment-shaped. Booked slots, service calls, consultations, estimates. If the value of the business is a calendar, an AI receptionist has a number attached to it. If they sell retail off a shelf, the arithmetic gets vague and vague loses.

Do they already pay for software. A booking widget, a CRM logo in the footer, a chat bubble, a review-request text you got after a past visit. A business paying $200 a month for scheduling software has already decided that software is a normal expense. You are proposing a line item, not a new category of spending.

Three of four is a call. Two of four is a nurture. One is a no.

Why "we already use AI" is a buying signal

This is the objection that throws people, and it should not.

Goldman Sachs surveyed 1,256 owners in its 10,000 Small Businesses program between January 27 and February 4, 2026. Seventy-six percent said they currently use AI. Only 14 percent said AI is fully embedded in their core operations. Another 73 percent said they would benefit from more training and implementation help. Read those three numbers together and the picture is a market that has tried AI, likes it, and has not wired it into anything that runs without a human present.

The spending data says the same thing in dollars. JPMorganChase found median monthly AI spend fell from roughly $80 in 2022 to about $30 by 2025, and the share of AI users in the lowest tier, up to $40 a month, grew from about 38 percent in 2019 to roughly 63 percent in 2025. Most small businesses are paying for a seat on a chat tool. Nobody is paying for a system that answers the phone at 9pm and books the job.

One fair caveat: the Goldman sample is self-selected, owners enrolled in a business education program, so they skew more deliberate than the average shop on a strip mall. The direction still holds against the transaction data, and the direction is what you need.

So when a prospect says they already use AI, the follow-up is not a defense of your product. It is one question: what does it do while you are asleep? Almost always the answer is nothing, and now you are describing a gap they discovered themselves.

Which trades take longer, and why you might still want them

The report's own framing is that "knowledge-intensive industries show significantly higher adoption than labor-intensive sectors." Information, professional services and educational services lead. Construction and transportation trail badly.

That looks like bad news for anyone selling into the trades. Be clear-eyed about it. An HVAC owner has seen fewer AI tools than a law firm partner has, so more of the first call goes to explaining what this even is. The compensation is that nobody else has sold him anything yet, and the pain is louder. A missed call at a law firm is a case someone follows up on Monday. A missed call at a roofing company during a hailstorm is gone in four minutes.

So sequence it. Low-adoption trades need a demo they can touch on the first call, because the concept does not land by description. High-adoption verticals need a reason yours beats the three tools they already pay for. Same product, different opening minute. The demo call playbook covers handing over the phone.

Headcount is a threshold, not a slope

The Census Bureau's Business Trends and Outlook Survey put national AI use at 19.8 percent as of May 3, 2026. Firms with 250 or more employees sat at 37 percent. Firms with four or fewer employees came in under 20 percent. The line that matters for your prospecting is this one: between December 2025 and May 2026, AI use rose among firms with at least 20 employees and did not change significantly among firms with fewer than 20.

Two readings, both useful. For the easiest sale, go up-market, because the 20-plus segment is already moving. For the emptiest room, stay under 20 employees, where the market has been flat for six months and no competitor is bothering. I prefer the second and I will defend it, but it means creating demand rather than catching it, which shows up as longer first calls and more no-shows. Price and pace accordingly. There is more on the flat line itself in the small business AI adoption gap.

Who I disqualify, and what it costs me

Most qualification advice skips this part. The hard nos:

Businesses with no phone problem. If the owner answers on the second ring at 7pm and loves it, an AI receptionist is a solution to a problem they do not have. You can sell them the website and chat piece, but do not lead with voice.

Pure product retail with no bookings. Census data puts retail trade around 14 percent AI use, below the national average, and the reason is structural rather than stubbornness. There is no appointment to protect.

Franchise locations without buying authority. The operator will love the demo, then discover vendor decisions happen at corporate. Two weeks gone.

Anyone whose real complaint is lead volume. An AI receptionist converts the calls you already get. It does not manufacture calls. If the phone rings four times a week, they need marketing first, and saying so out loud buys you more credibility than a stretched pitch.

Businesses that cannot absorb more jobs. A two-man crew booked eight weeks out does not want a better booking rate. Sell them the after-hours filter, or wait until they hire.

I lose real deals to that list every month. The tradeoff is that the deals I do close stay closed, which matters more when your revenue is monthly.

Where Lumina Sales fits in this

Qualification is cheap when the demo is cheap, which is most of why our model looks the way it does. You build a live demo site by picking a trade, typing the prospect's business name, and arranging the nodes you want. The 100-plus vertical packs carry real services, price bands and objections, so a low-adoption trade hears its own language instead of a template with the industry left blank.

Platform access is $49 a month with the limits printed next to it: 100 live voice minutes, 500 chat messages, 50 website audits, one published demo at a time, five minutes maximum per call. Client builds are quoted on a 30-minute call at a wholesale setup fee and monthly, and you set retail. Full numbers sit in pricing. One published demo at a time is the real constraint on a heavy prospecting week. Plan your list around it.

FAQ

Is number of employees really a better filter than revenue?

On the JPMorganChase payment data, yes. Employer firms with revenue under $250,000 adopted AI at 27.6 percent while nonemployer firms above $250,000 sat at 19 percent. Revenue still matters for what you can charge. It just does not predict whether they will say yes.

How do I find out headcount before I call?

The website is right most of the time: team pages, plural job titles, hiring links. Failing that, ask on the call. It is a normal question.

Should I disqualify solo operators entirely?

No, rank them last. Nonemployer adoption still reached 15.3 percent by December 2025, so the buyers exist. They take more calls to find, and a solo owner is likelier to try assembling something himself.

What if a prospect uses free AI tools already?

That counts in your favor. The SBE Council's March 2026 technology survey found 82 percent of small business employers had adopted at least one AI tool, with a median of five tools in use. Five disconnected tools is a workflow problem, and a single system that answers, books and follows up is the answer to it.

Does any of this change how I price?

Longer sales cycles in low-adoption trades argue for higher monthly pricing, not lower, because you are earning the education. How much to charge for an AI buildout has the math.

Do this with your current list

Open your prospect sheet and add one column: employees, yes or no. Sort by it. Call the yes column at 6:40pm this week and log what the phone does. The list gets shorter and the hit rate gets better, which is the only trade worth making when your time is the scarce input.

If you want the demo half of this handled, start an agency workspace and build one for the best-qualified name on the sheet before you call them.