Research
We track how AI actually shows up in small and mid-sized organizations, the kind with ten to two hundred people and little or no technology staff. Where the tools arrive before the guidance does. And awareness is rarely the problem, the translation is.
The gap
Under 20%
The share of firms with four or fewer employees that used AI in the past two weeks, against 37 percent of firms with 250 or more. The Business Trends and Outlook Survey asks about real use, not plans.
Why it matters
The smallest firms are the furthest behind, and they are the ones we work with. This is the federal baseline, measured by what businesses actually do.
About 18%
Of firms had adopted AI by the end of 2025, once the Fed reconciles three national surveys. Adoption climbs steeply with company size.
Why it matters
A sober read from the central bank. AI is not yet everywhere on Main Street, whatever the louder claims say.
76% use it, 14% live on it
Owners who use AI, against owners who have built it into core operations. From a survey of 1,256 owners by Babson College and David Binder Research.
Why it matters
The gap in one line. Most owners have the tools in hand. Few have woven them into how the business actually runs. The distance between the two is the work.
26.1% vs 15.3%
AI adoption among employer firms versus nonemployers by December 2025, measured from real payments across millions of business accounts. The gap has nearly doubled since early 2023.
Why it matters
Measured by where the money actually goes, not by what owners say. The firms with staff are pulling ahead, and the rest are falling further behind.
95%
Of enterprise AI pilots delivered no measurable return. MIT puts the cause on the gap between tools and real work, not on the technology.
Why it matters
Even large companies with budgets stall here. The lesson holds at any size. A tool you have not connected to the work produces nothing. Translation comes first.
Where it stalls
Nearly 82%
Of businesses with fewer than five employees said AI was not relevant to them, in the Census AI supplement the report draws on. The SBA also notes that federal AI resources are not centralized enough to help.
Why it matters
The smallest owners do not see how AI applies to their work. That is a translation problem, named by the government’s own small business watchdog.
Many small business owners seem eager to deploy AI to support their enterprises, but aren’t sure where to start.
Sarah Simms of the Federal Reserve Bank of San Francisco, drawing on the Small Business Credit Survey.
Why it matters
A named Fed official saying plainly what we hear in every first conversation. The want is there. The path is not.
36%
Of more than 3,000 workers said they have the training and resources to use AI, down from 45 percent a year earlier, even as use rose. Survey by Jobs for the Future, republished by PSHRA.
Why it matters
Tools without training do not stick. Small teams feel that gap hardest, because there is no one down the hall to ask.
Industry by industry
25% use it, 43% distrust it
Of residential contractors use AI in a real way, while 43 percent do not trust its output. ServiceTitan, which sells software to the trades, surveyed 1,000 contractors through the independent firm Thrive Analytics.
Why it matters
HVAC, plumbing, and electrical owners already wear every hat. They do not need more noise. They need someone to turn it into a next step.
63% explore, 16% use
Finance professionals are exploring AI, but only 16 percent actually use it. From a survey of 128 professionals by Accounting Seed, an accounting-software vendor, reported by CFO.com.
Why it matters
In the back office, trust and accuracy decide everything. Teams stall because they cannot tell which tools are credible. That judgment is exactly what an assessment provides.
92% use it, 7% gain from it
Of nonprofits use AI in some form, but only 7 percent report major gains, often because the know-how sits with one person. Survey of 346 nonprofits by Virtuous, a nonprofit software vendor.
Why it matters
Almost everyone is experimenting. Almost no one is getting real return, because scattered experiments are not a system. Mission teams need those experiments turned into shared workflows.
On the ground
Everybody wants to use AI, but it’s not like a blanket: here is your AI, use this for your business.
Priya Biswas, executive director of the Kootenay Association for Science and Technology, whose provincially funded TechEdge program helps rural owners assess which tools fit before they buy.
Why it matters
The translation gap, described from the ground. The work starts before a business ever picks a tool. That first read is an assessment, and it is where any honest AI effort begins.
I think people are still trying to wrap their brains around it.
Laurene Rodriguez, CEO of Mariposa Marketing. One local owner took a six-week prompting course at the community college just to get started.
Why it matters
The appetite is real, and so is the effort it takes to get going. People are teaching themselves because no one has handed them the path.
About 12%
Of Canadian businesses reported using AI in the second quarter of 2025, per Statistics Canada. Canada’s small business bank is now pairing 500 million dollars in financing with human advisors, an admission that money alone does not close the gap.
Why it matters
Even a national bank concluded that capital is not enough. Owners need a person to help them see where AI actually fits. That is the whole idea.
A note on sources. Every figure here was checked against the original report. The government and institutional sources carry the most weight, and where a survey was run by a company that sells the tools, we say so in the line beneath it. If you have seen research worth adding, send it along.
Just a plain-language look at what's possible for your organization. If it fits, we build it together.