Bank Data, Not Surveys: 1 in 6 Small Businesses Now Pays for AI — and Solo Operators Are Falling Behind

Illustration of an ascending bar chart made of building blocks on a dark navy background, with a rising line passing glowing milestone nodes, a storefront icon, and an hourglass icon beside the base of the chart, electric-blue accents

Every few weeks a survey announces that small businesses are "adopting AI" at some impressive rate. Surveys measure intentions. In April 2026, the JPMorganChase Institute published something better: a report that tracked actual bank payments to AI services across 4.6 million small businesses, from 2019 through 2025, using de-identified Chase Business Banking transaction data. No self-reporting, no aspirational answers — just who actually paid, when, and how much.

The headline number: by the end of 2025, 17.7 percent of small businesses had paid for an AI service, up from 5.2 percent just two years earlier — the adopter base more than tripled in 24 months. But the number that should stop a small-business owner isn't the growth rate. It's the gap underneath it: businesses with employees adopted AI at nearly double the rate of solo operators with the exact same revenue. The report's own interpretation of that gap has nothing to do with money — and everything to do with bandwidth.

What the Bank Data Actually Shows

Three findings stand out from the transaction data, and together they describe a market that changed shape in two years.

Adoption accelerated roughly thirteen-fold across cohorts. Firms that started in 2019 needed 77 months — over six years — before 10 percent of that cohort had ever paid for an AI tool. The 2023 cohort got there in 23 months. The 2024 cohort in 15. The 2025 cohort reached 10 percent adoption in about six months. New businesses now sign up for AI the way they sign up for internet service: 6.5 percent of firms formed in 2025 paid for an AI tool in their very first month, more than five times the rate of the 2019 cohort.

Spending flattened because entry got cheap, not because commitment fell. Median monthly AI spending peaked around $80 in 2022 and fell to roughly $28–30 by 2025. That decline is a composition effect, not a retreat: entry-level generative AI subscriptions at $20–30 per month created a new on-ramp, and by 2025, 63 percent of AI-paying small businesses sat in the $1–40/month tier (up from 38 percent in 2019). Meanwhile, businesses that adopted early kept deepening — the 2019 cohort grew from about $50 to roughly $90 a month over six years, an 80 percent increase — while 2024 adopters entered around $21 and stayed nearly flat.

Consistent use replaced experimenting. Consistent AI payers outnumbered sporadic ones in every year measured, and by 2025 the ratio hit its highest level in the series. In 2019, 89 percent of AI-paying firms paid for exactly one service; by 2025, 28 percent paid for two or more, with generative AI the dominant paid category in every single industry the report tracked.

The Real Divide Isn't Revenue — It's Bandwidth

Here is the finding that matters most if you run a small operation. By December 2025, 26.1 percent of employer firms had adopted AI versus 15.3 percent of nonemployer firms — a gap that widened from 5.6 percentage points in January 2023 to 10.8 points at the end of 2025. That alone could just mean bigger businesses have bigger budgets. So the researchers cut the data by revenue, and the budget explanation fell apart.

Among firms earning under $250,000 a year, employers still adopted at 27.6 percent versus 14.3 percent for solo operators — nearly a two-to-one gap at identical revenue levels. Small employers even out-adopted large nonemployers, 27.6 percent to 19 percent. The report's conclusion: the divide reflects organizational capacity and human capital — the time, skills, and bandwidth to implement and integrate — rather than the financial capacity to afford AI services. A solo owner is the marketing department, the ops department, and the help desk; "learn and deploy a new AI stack" loses every scheduling battle. An NFIB survey cited in the report backs the familiarity gap: only 48 percent of nonemployers reported being familiar with AI to some degree, versus 82 percent of firms with 50 or more employees.

Read that correctly and it is both a warning and an opening. The warning: waiting doesn't make this cheaper or easier — the gap between plugged-in and unplanned businesses has widened every year since 2023. The opening: in industries where most competitors are solo or micro-operators, almost nobody has operational AI yet, and the first business in a local niche to actually deploy it gets the whole advantage.

The Temecula and Murrieta Angle

Our local economy runs on exactly the industries the data flags as late adopters. Accommodation and food services, retail, arts and entertainment, construction — all sit well below the 39.3 percent adoption of the information sector, with construction at 8.9 percent and transportation at 5.4 percent. And where those industries do pay for AI, the spending is almost entirely entry-level generative tools — 62 percent of food-service AI spending went to general-purpose generative AI, the biggest single share of any industry.

Translation for a Temecula tasting room, a Murrieta contractor, or a valley restaurant: your competitors' "AI strategy" is, at best, a shared ChatGPT subscription someone uses to write posts. The operational layer — answering every call, following up on every quote request within minutes, booking while you sleep — is completely unclaimed locally. That is the layer that decides which business gets the job when three quotes come in on a Tuesday night, and the bank data says your market won't close that gap on its own for years.

What to Do If You're the Bandwidth-Constrained Business

The report accidentally wrote the playbook for solo and micro operators, because it documented what doesn't work: the $20–30 monthly subscription tier, adopted alone, produced years of nearly flat engagement. The businesses that got value are the ones whose spending grew — the ones that integrated AI into actual operations. Four moves follow:

Don't buy tools; buy implemented outcomes. The data says the constraint is implementation capacity, not access. A tool you never set up is a flat line. A system someone sets up, trains, and maintains for you is the employer-firm advantage outsourced. That is literally the service model we built — flat monthly plans, no contracts, where the setup and maintenance are ours.
Start where revenue leaks, not where AI is flashy. Missed calls and slow lead follow-up are the two most measurable leaks in a service business, and both are solved by the same pattern. Our AI voice agent guide covers the realistic setup and what it costs.
If you do run tools yourself, meter the spend. Self-managed AI costs drift — token-based pricing especially. Our hidden bill guide shows where the surprise charges live and how to cap them.
Pair AI with the labor you can't hire. If you've tried and failed to hire help this year, automation is the bridge, not a betrayal of personal service. See how the math works for local service businesses.

The Bottom Line

One in six American small businesses now pays for AI, and the number is compounding — but the median spend is $28 a month, and the fastest-growing divide in the data is not between big budgets and small ones. It is between businesses that have someone to implement AI and businesses that don't. If you have a team, assign an owner. If you are the team, that is exactly the kind of work you should never do yourself: it is buyable, it is flat-fee, and it starts returning the month it goes live.

PepeWebTech deploys exactly this operational layer for small businesses in Temecula and Murrieta — chat and phone agents that answer and book around the clock, follow-up systems that respond in minutes, and websites built to convert. If the bank data describes your situation, book a free 15-minute call or see the plans.