If you run a small business, you have probably been told two contradictory things about AI this year. One camp says it will change everything. The other says it is an expensive toy for companies with IT departments. Neither tells you the only thing you actually want to know: what does it cost someone like me, and what do I get back?
This year there is finally a decent answer, from owners rather than vendors.
The Small Business & Entrepreneurship Council surveyed 517 owners of businesses with 2 to 99 employees in February. The typical owner spends a median of $2,200 a year on AI, about $183 a month. In return, the median owner says they personally get back five hours a week, and their staff together get back 11.5 hours a week. Two-thirds report revenue gains they attribute to AI, and 22% say those gains were above 10%.
That sounds like a no-brainer. It mostly is. But there is a catch hidden in "hours saved", and it is the difference between the owners who ended the year with more money and the ones who ended it with slightly more free afternoons.
What owners actually spend
The fear that AI means enterprise contracts turns out to be wrong for most small businesses. Here is the real spread.
| Annual spend on AI | Share of small business owners |
|---|---|
| $1,000 or less | 42% |
| $1,000 to $25,000 | nearly half |
| More than $25,000 | 8% |
| Median | $2,200 a year |
Homebase, which surveyed 750 owners and managers of businesses with hourly teams in a report published on 22 September, found the same shape from a different angle: among adopters, the most common spend is $41 to $150 a month, and 13% pay nothing extra at all because the AI sits inside software they already use.
So for most owners this is not a capital decision. It is a subscription, roughly what you might spend on accounting software or a second phone line.
What they get back
| What owners report | Median or share |
| Owner's own time saved | 5 hours a week |
| Staff time saved, across the business | 11.5 hours a week |
| Report revenue gains attributed to AI | two-thirds |
| Report revenue gains above 10% | 22% |
| Report AI saved them time at all (Homebase) | 87% |
| Owners using AI for scheduling and payroll: estimated saving | 7.7 hours and $343 a month, about $4,100 a year |
Owners also told SBE Council what they use it for, and it is unglamorous: general business research (37%), writing content (30%), and marketing automation, project management and social media at about a quarter each. Homebase's adopters lean even more practical: analysing data (47%), payroll and timekeeping (45%), inventory (44%) and paperwork (43%).
None of that is a robot running your shop. It is the boring work that eats an owner's evenings, done faster.
The owner's math, worked through
Put the two medians together for a single, typical owner.
| Per year | |
| Spent on AI | $2,200 |
| Owner hours saved (5 a week, 50 weeks) | 250 hours |
| Value of those hours at $40 an hour | $10,000 |
| Value at $75 an hour | $18,750 |
| Staff hours saved (11.5 a week, 50 weeks) | 575 hours |
On paper, that is a return of roughly four and a half to eight and a half times the spend from the owner's time alone, before counting staff hours or any revenue gain. Even if you cut the self-reported time in half to allow for optimism, it still pays comfortably.
Here is the catch, and it is the most important line in this article: an hour saved is not a dollar earned. The $10,000 is only real if you spend those 250 hours on something that makes or keeps money. If the saved time goes into answering slightly more emails, or leaving at five instead of six, the AI has made your life better, which is worth something, but it has not made your business richer.
What the owners who gained did with the time
This is where the surveys get genuinely useful, because they asked.
Among SBE Council's owners, 44% say they reinvested the efficiency into better innovation or customer engagement, 35% into technology or competitiveness, and 34% into additional growth opportunities. In Homebase's survey, of the owners whose time was freed, 57% put it toward growing the business and 53% toward areas they had been neglecting.
Read those two lists together and a pattern appears. The owners reporting revenue gains are the ones who decided in advance what the hours were for. Not "I'll have more time" but "I'll use Tuesday afternoons to call lapsed customers", or "I'll finally build the quote template that loses us jobs", or "I'll answer every enquiry within the hour instead of the next day".
That last one is a good example of why this works. A business that replies to leads in minutes rather than hours wins noticeably more of them, and AI makes fast replies cheap. I worked through the numbers for that one specific use in AI lead response for small businesses. It is the clearest case where saved time turns straight into revenue, because the time saved is the speed the customer notices.
The mistake that wastes the $2,200
The most common way small businesses waste AI money is buying a tool before deciding what problem it solves. Homebase found that among owners who have not adopted, 62% are still interested, but only 20% have looked into specific tools, and 18% did not know that AI tools for scheduling, payroll or hiring even existed.
The opposite failure is just as common: buying five tools (the median owner already uses five) and wiring none of them into how the business actually runs. The result is subscriptions that each save a few minutes and nothing that changes the week.
The fix is the same in both cases and costs nothing: write down the three tasks that eat the most of your week, pick one, and fix that one properly before buying anything else. That is the whole argument of fix the workflow before you buy another AI tool, and the step-by-step version is the implementation playbook for small businesses.
Will it cost jobs in your business?
Owners worry about this more than they say, partly for their staff and partly because hiring is hard enough already. The data is calmer than the headlines.
| Effect on employment, SBE Council | Share of owners |
| No overall change | 48% |
| Net job creation | 19% |
| Net reductions | 9% |
| See AI as complementary to workers | 52% |
| See it mainly as a substitute | 8% |
Twice as many small businesses report adding jobs as cutting them. For a business with ten people, AI mostly shows up as fewer hours lost to admin, not fewer people. The staff time saved, 11.5 hours a week across the median business, is a little over a quarter of one full-time job spread across everyone, which is usually absorbed as less overtime and fewer evenings rather than a redundancy.
Who is getting left behind
Adoption is not even, and the gap tracks money.
Homebase found adoption rises from 57% at businesses with 5 to 9 employees to 84% at 50 to 99, and from 54% at businesses under $500,000 in revenue to 81% above $5 million. Restaurants and food service had the lowest adoption of the industries it listed, at 65%. Among non-adopters, the top reasons were data privacy and security (38%), fear of mistakes (31%) and not trusting AI (27%).
Those are reasonable worries, and they are also solvable ones. Owners said what would change their minds: knowing their data is secure, evidence of cost savings, and the ability for a person to override the AI's decisions. If you are in that group, those three are exactly the questions to ask any tool before you pay for it. Restaurant owners specifically will find the practical cases in AI for restaurants.
What this means for you
If you have not started, the honest starting budget is small. A single subscription in the $40 to $150 a month range, pointed at your single most painful task, is how most owners who now see results began. Do not buy five things. Buy one, and decide before you start what you will do with the hours it gives back.
If you already pay for AI but cannot point to a result, you are probably in the large group that saved time without redirecting it. Pick one revenue task, such as follow-ups, quotes or lapsed customers, and book the saved hours into your calendar for it. Give it a month, then compare the revenue line, not the time-saved line.
If you are an owner deciding whether this is a fad, look at the investment intentions. 62% of SBE Council's owners expect to increase AI spending next year and only 3% expect to cut it. That is not what a fad looks like among people spending their own money.
If you sell services to small businesses, this is a market telling you exactly what it wants: one problem solved properly, at a monthly price, with a person who can override the machine. That is a business, and it is one of the routes in the nine ways people actually make money with AI.
The honest take
The striking thing about the small-business numbers is how ordinary they are, in the best sense. The typical owner spends about what a phone contract costs, gets back an afternoon a week, and in two cases out of three sees revenue move. It is not a revolution. It is a modest, reliable improvement that most owners can afford, which is rarer and more useful than a revolution.
But the surveys also show where the value leaks. Saved time is the easiest thing in the world to fritter away, and the owners who end the year richer are not the ones with the best tools. They are the ones who decided, before they switched anything on, which customer, which quote or which follow-up those hours were going to.
So here is the question to answer before you renew any AI subscription: if it gave you back five hours last week, can you name what those five hours earned?
Frequently asked questions
How much does AI cost a small business?
The median small business owner spends about $2,200 a year on AI, roughly $183 a month, according to a February 2026 survey of 517 owners of businesses with 2 to 99 employees. 42% spend $1,000 a year or less, and among adopters in a separate survey the most common spend was $41 to $150 a month.
How much time does AI save a small business owner?
A median of five hours a week for the owner personally and 11.5 hours a week across staff. Owners using AI for both scheduling and payroll estimated saving 7.7 hours and $343 a month.
Does AI actually increase small business revenue?
Two-thirds of owners report revenue gains they attribute to AI, and 22% report gains above 10%. These are self-reported. The owners reporting gains tend to be the ones who redirected saved time into sales and customer work.
Will AI make me cut staff?
For most small businesses, no. 48% report no change in employment, 19% report net job creation and 9% report net reductions. Most owners see AI as complementary to their staff.
Where should a small business start with AI?
With one task, not one tool. Write down the three tasks that take most of your week, pick the one closest to revenue, and solve that properly before buying anything else.
Sources: SBE Council: Small Business Technology Use Survey, March 2026; Homebase: 2026 Main Street AI Gap Report, 22 September 2026.
