What Gallup's AI adoption survey means for small businesses

You paste a client email into an AI chat and ask it to sound more professional. Back it comes, opening with "I hope this message finds you well" (an example written for this post, though you may recognise it).

That is most of what you use AI for, and it saves a few minutes.

Gallup's May 2026 AI adoption figures for US workers suggest that is where most people start.

So what does the Gallup AI adoption survey mean for a coach, consultant or freelancer? Writing and research are the top uses, but the people who report the biggest productivity gains use it for specific tasks, and in more ways. The practical move is to add one specific job, not more of the same.

What did Gallup find about AI at work in the US?

Gallup's report, Organizational AI Adoption Jumps Six Points, is based on its quarterly survey of US employees, run in May 2026 and published in July. The headline findings, in Gallup's words:

• "Forty-seven percent of U.S. employees now say their organization has integrated AI tools to improve productivity, efficiency or quality, up from 41% in the last quarter."
• "More than half of U.S. workers (52%) now use AI in their role, with 30% using it frequently (a few times a week or more)."
• "Fifteen percent use it daily."
• "One in five employees remain unsure as to whether their organization has integrated AI tools, unchanged from prior quarters."

Two things are worth knowing before you lean on those numbers.

Gallup's AI indicator page explains that its total counts employees who used AI at work "a few times a year or more", so the frequent group is the better picture of real habits. And these are US employees working for organisations, not UK sole traders, so treat it as a view of how people at work use AI.

As for the one in five who aren't sure whether their organisation uses AI, in a business of one that question has a short answer.

What UK figures show

UK official figures measure something different: whether businesses use AI, not whether workers do. So read them alongside Gallup's US figures, not against them.

The Office for National Statistics surveys UK private sector businesses. In its 8 October 2026 bulletin, "around 3 in 10 (30%) of all businesses reported using at least one type of artificial intelligence (AI) technology in September 2026", and "only 29% of businesses with fewer than 10 employees reported using AI technologies." The ONS calls these official statistics in development.

The UK government's AI adoption survey 2026, published on 29 October 2026, interviewed 3,503 businesses between January and April 2026, against a target of "3,500 UK private sector businesses with at least 5 employees". "In 2026 32% reported using one or more AI technologies", including "27% of micro businesses (5 to 9 employees)." Its smallest band starts at five people, so a business of one isn't counted.

Why does most AI use stop at writing and research?

Gallup found that "among AI users, the most common uses of AI are writing and editing (51%), search or research (49%), and general assistance or problem-solving (39%)." It describes this as "knowledge support". More specific uses are rarer: "coding assistance and automation, each cited by 16% of AI users."

The reason is easy to guess, though this part is our reading rather than Gallup's. A chat box looks like a place to write and ask questions, so that is what most of us do with it. A specific task needs you to decide what the job is first: what goes in, what should come out and what done looks like.

That matters because, as Gallup puts it, "the AI applications that employees use most often are not the ones most strongly associated with productivity improvements."

Where people reported the biggest gains

The highest ratings came from people using AI for defined jobs. For coding assistance and for automation, "more than three-fourths of workers who use AI in each of these ways (77%) say AI has had an extremely or somewhat positive effect on their productivity." Ratings were "nearly as high among employees using AI for presentation or slide deck creation (76%) and for data science or analytics (75%)."

Writing and research still help.

Gallup reports that "sixty-eight percent of employees who use AI for writing and editing say it has improved their productivity, as do 65% of those using it for search or research." So your email tidying isn't wasted, just not where people report the most value. These are people's own ratings, not measured output.

What using AI in more ways does and doesn't show

Gallup also looked at variety:

"Among employees using AI for one or two purposes, 45% say it has had a somewhat or extremely positive impact on their productivity. That rises to 66% among those using it for three or four purposes, to 78% among those using it for five or six purposes, and to 90% among those using it for seven or more purposes."

Gallup is careful here, and so should we be: "This relationship alone does not prove that adding more AI applications will have a greater impact." It adds that "employees who see more value in AI may be more likely to find additional ways to use it."

So the lesson isn't to collect uses like loyalty stamps. People who get real value tend to have found several specific jobs for AI, one at a time.

Add one job with a clear start and finish

The fix is small. Pick one task from your week that has a clear input, a clear output and a format you already use. For a coach, consultant or freelancer, that might be:

• turning your notes from a client call into a session summary in your usual layout
• drafting a slide outline from a proposal you've already written
• sorting a month of enquiry emails into common questions
• checking a spreadsheet of invoices for missing dates or duplicates

These sit close to the kinds of task Gallup's report links with higher ratings, such as presentation creation, analytics and automation, but they are our examples, not Gallup's categories.

The test is simple: if you can't say what done looks like, it isn't a specific job yet.

Before and after: the same afternoon, two ways

Before, the request most of us type:

Make this email sound more professional.

After, a specific job with an input, an output and a format:

Here are my rough notes from a discovery call with a new coaching client. Turn them into a session summary using these headings: what they want, what's in the way, what we agreed, next steps. Keep it under a page, use their words where you can, and list anything I didn't capture that I should ask about next time.

A reply might look like this (an example written for this post):

What they want: to stop working weekends by the spring. What's in the way: every enquiry comes to them directly. What we agreed: a shared inbox trial. Next steps: they send last month's enquiries. Not captured: who else answers the phone.

The first request polishes something you'd already done. The second does a job you would otherwise have done yourself, and it tells you what's missing.

What this won't fix

Gallup's figures can't tell you which task will pay off in your business. They are self-reported, from US employees, and the link between variety and productivity is a pattern, not proof.

A specific job still needs checking: AI can tidy up a summary and quietly drop the one detail that mattered. And if your process is muddled, AI will follow the muddle faithfully.

One thing to do today

Look back at the last few times you used AI. If they were all writing or looking things up, pick one task from this week with a clear start and finish and give it a proper brief.

If you're not sure where to start, the AI Readiness Check is a short check of where AI could help your business most, with no account needed. When you want to go further, the course library sets out the courses in the order they build on each other.