Ignited Labs
Ignited Labs · Research

AI Adoption in Small Business: The Gap Between Use and Business Value

AI adoption in small business is no longer the question. It's settled. The question is why almost no one manages to get value out of it, and why the smallest businesses in particular stay out of the picture. That's what 133 verified statistics show.

133Unique statistics
17Source organizations
2023-2026Data range
June 2026Last updated
01 - Bottom Line

Everyone adopts. Almost no one profits.

Three numbers explain all the rest. In each one there's a gap between what businesses say and what they actually hold. That gap - not price, not access - is the real barrier.

88%→6%
of organizations use AI regularly, versus only 6% considered to be extracting real value from it Gap: 82 percentage points McKinsey, State of AI 2025
94%→27%
say connected data and processes are critical to AI success, versus 27% who actually built that connection Gap: 67 percentage points HBR Analytic Services, Dec 2025
96%→82%
of small business owners plan to adopt new technologies including AI, versus 82% of businesses with fewer than 5 employees who say AI is simply not relevant to them Both numbers are true. That's the whole story. U.S. Chamber 2025 · SBA Advocacy, Sep 2025

And one number that takes the financial argument off the table entirely: a ChatGPT Plus and Claude Pro subscription together costs around $40 a month. A freelancer or marketing agency costs between $500 and $3,000 a month. The capability is no longer expensive. Knowing what to do with it is what's missing.

02 - Adoption Is Settled

It happened faster than electricity

Generative AI use among small businesses in the U.S. nearly doubled, twice, in two years. By 2025 the majority had already crossed the line. Whoever is waiting for the "adoption wave" is waiting for something that already happened.

Generative AI use among small businesses in the U.S.

Percent of small businesses reporting generative AI use, by annual survey

2023
23%
2024
40%
2025
58%

The scale runs up to 58%, the highest value in the series. U.S. Chamber of Commerce, 2024 & 2025

Data as a table
YearGenerative AI useSource
202323%U.S. Chamber of Commerce, 2024
202440%U.S. Chamber of Commerce, 2024
202558%U.S. Chamber of Commerce, 2025
AI reached 45% usage among American workers 13 times faster than the equivalent technology in 2019. Electricity took more than thirty years to reach the same point. JPMorgan Chase Institute, Apr 2026
03 - The Pattern That Repeats

Four gaps, the exact same shape

This isn't one random gap. In every survey, with every measuring body, across every breakdown, the same picture repeats: what businesses agree matters sits in the sky, and what they actually built sits on the floor.

What they say / agree matters What they actually built or do

Connected data and processes are critical to AI success · vs. who actually connected them · gap 67 pts

Say
94%
Did in practice
27%

Use AI regularly · vs. who extracts value from it · gap 82 pts

Say
88%
Did in practice
6%

Personalization improves leads and sales · vs. who personalizes based on data · gap 80 pts

Say
93%
Did in practice
13%

AI is essential to stay competitive · vs. who integrated it into daily operations · gap 57 pts

Say
82%
Did in practice
25%

Each pair comes from the same survey, so the comparison is only ever between the same respondents.

Data as a table
What was measuredSayDoGapSource
Connected data & processes94%27%67 ptsHBR Analytic Services, Dec 2025
Regular use vs. extracting value88%6%82 ptsMcKinsey, State of AI 2025
Personalization93%13%80 ptsHubSpot, State of Marketing 2026
Essential vs. daily integration82%25%57 ptsReimagine Main Street / PayPal, Jun 2025
04 - The Breakdown That Matters

The smaller the business, the wider the gap

And here's the number that turns this from a general problem into a specific one: the gap between a sole proprietor and a business with employees doubled in two years, from 5.4 percentage points to 10.8. It isn't closing. It's opening.

And it isn't about money. The JPMorgan Chase Institute checked and found that businesses with employees are ahead of businesses without employees at every income level. The difference isn't budget - it's spare capacity: time, skill and attention.

The very smallest businesses The larger comparison group

AI adoption rate · sole proprietor vs. business with employees · gap 10.8 pts

Sole proprietor
15.3%
With employees
26.1%

See AI as important to the business · 1-9 employees vs. 50+ employees · gap 32 pts

1-9 employees
37%
50+ employees
69%

Basic AI familiarity · sole proprietor vs. 50+ employees · gap 34 pts

Sole proprietor
48%
50+ employees
82%

Three different surveys for the three rows, so don't compare across rows - only within each row.

Data as a table
What was measuredThe smallThe largeGapSource
AI adoption15.3%26.1%10.8 ptsJPMorgan Chase Institute, Dec 2025
AI seen as important37%69%32 ptsNFIB Survey, 2025
AI familiarity48%82%34 ptsU.S. Chamber via JPMorgan
Adoption gap, 20235.4 pts→ 10.8 pts by 2025JPMorgan Chase Institute

What the small say about themselves

82% of businesses with fewer than 5 employees say AI is not relevant to them SBA Office of Advocacy, Sep 2025
25% say there's nothing in the market that would meaningfully improve their business NFIB Survey, 2025
41% of small business employees say it's unclear how AI touches them U.S. Chamber Foundation, Jun 2026
50% of small businesses that do use AI do so with zero dedicated investment SBA Office of Advocacy, 2025
05 - By Industry

The gap isn't technological. It's translational.

The exact same technology, the same price, the same access. And yet one industry adopts at 13 times the rate of another. What separates them isn't the tool - it's how easy it is to see why the tool matters for everyday work.

AI usage rate by industry

Industries are categories with no natural order, so they're all the same color. Sorted by size only.

Professional services
18.2%
Information & communication
18.1%
Manufacturing & industry
~12%
Retail
~4%
Hospitality & restaurants
2.9%
Construction & agriculture
1.4%
0%5%10%15%20%

The two leading industries are highlighted. Manufacturing and retail values are marked ~ because the source itself gives them as a rounded estimate, not an exact figure. Census BTOS · Brookings · McElheran et al., via JPMorgan

Data as a table
IndustryAI useSource
Professional services18.2%Brookings via JPMorgan, 2025
Information & communication18.1%Census BTOS via JPMorgan
Manufacturing & industry~12%McElheran et al. via JPMorgan, 2024
Retail~4%McElheran et al. via JPMorgan, 2024
Hospitality & restaurants2.9%Brookings via JPMorgan, 2025
Construction & agriculture1.4%Census Bureau via JPMorgan
06 - Why The Gap Exists

What's missing isn't the model. It's the context.

HubSpot asked 1,505 marketers which data about their customers they even have on hand. The answer forms a ladder: the more valuable the data, the fewer people hold it. At the bottom of the ladder, the most valuable information of all - what bothers the customer - sits with one in six.

This explains the gap we just saw. You can't get an answer from AI that fits your business if your business isn't written down anywhere the AI can read.

What customer data marketers actually have access to

Ordered from most available to rarest. Percent of 1,505 marketers.

Purchase habits
44%
Basic demographics
39%
Purchase history with me
31%
What they're interested in buying
27%
What they expect from service
17%
Their pains and challenges
16%
0%10%20%30%40%50%

The top-to-bottom order is depth of context. The scale goes up to 50%, meaning no data point here is held by even half the market. HubSpot, State of Marketing 2026

Data as a table
Data typeHas access
Purchase habits44%
Basic demographics39%
Purchase history31%
Buying intent27%
Service expectations17%
Pains & challenges16%

All rows: HubSpot, "The State of Marketing 2026," a survey of 1,505 marketers, collected September 2025.

What this means for marketers

40% of teams never defined or wrote down anywhere what makes them different HubSpot, 2026
53% of marketers struggle to differentiate their content in an AI-saturated market HubSpot, 2026
71% say AI lets them produce far more content, while at the same time struggling to produce content that works HubSpot, 2026
61% agree that a clear point of view and distinct taste are a necessary component of an AI strategy HubSpot, 2026
07 - The Price

AI without context doesn't stay neutral. It's harmful.

This is the least-discussed point in the data. When AI runs without knowing the business, the result isn't "less useful." The result is work that needs fixing, and that costs real time. BetterUp Labs and Stanford's research has a name for it: workslop - output that looks fine and isn't.

41% received workslop from a colleague in the month before the survey BetterUp Labs / Stanford, HBR Sep 2025
1:56 hours and minutes it takes to fix and clean up each individual case BetterUp Labs / Stanford, Sep 2025
95% of organizations that tried AI report zero return MIT Media Lab, Aug 2025
19% less likely to reach the correct answer, when AI is used outside the domain where it's strong HBS Field Experiment

And against that, the same Harvard experiment shows what happens when AI is deployed inside the domain where it's strong: 12.2% more tasks, 25.1% faster, and quality 40% higher. Same tool, same people. The difference is whether someone knew how to aim it correctly.

The numbers don't argue about whether AI works. They show it works powerfully within certain boundaries, and does harm outside them. All the value depends on one question: who sets the boundaries.
08 - The Other Side

And for those who do close the gap, it works

For the picture to be complete: among the businesses that did reach orderly use, the numbers are consistent across completely different sources, and that's exactly what makes them credible.

91% of small and midsize businesses using AI say it increases revenue Salesforce SMB Trends, Dec 2024
58% of users save more than 20 hours a month - more than half a work week Thryv, Jul 2025
66% of users save between $500 and $2,000 a month Thryv, Jul 2025
82% of small businesses using AI reported growth in headcount U.S. Chamber of Commerce, 2025
5.6 hours per week saved on average per employee. For managers, 7.2; for other employees, 3.4 Business.com, Jan 2026
83% of growing businesses adopted AI, versus 55% of declining businesses Salesforce, Dec 2024

The last line is worth reading twice. AI adoption and growth show up together. The number doesn't say what causes what, but it does say the gap we measured on this page already shows up in the difference between growing businesses and businesses that aren't.

09 - What This Means

Four conclusions the data supports

Not forecasts, not product recommendations. Just what can be inferred directly from the numbers above.

88% → 6%

Adoption stopped being an advantage

When almost everyone uses it, usage itself no longer differentiates anyone. What differentiates is quality of execution, and that's exactly where 94% of the market hasn't built anything yet.

5.4 → 10.8

The window for the small is closing, not opening

The gap between sole proprietors and businesses with employees doubled in two years. If the trend continues, the smallest businesses aren't starting behind and catching up later - they're falling further away.

82% / 41%

The barrier is framing, not technology

82% of the tiniest businesses say "not relevant to me" and 41% of employees say "unclear how this touches me." Neither answer is about capability. Both are about the fact that no one translated the capability into their work.

16%

Context is the asset, not the tool

Only one in six knows what bothers their customers. Every business that does write this down somewhere AI can read it gets an advantage competitors can't out-feature.