How UK businesses really use AI in marketing: the government data

If you read the marketing press, you'd think every business in the country had rebuilt its marketing around AI. The most reliable UK data tells a different story, and it's more useful for anyone deciding what to do next.

In January 2026 the Department for Science, Innovation and Technology published its AI Adoption Research, based on interviews with 3,500 UK private sector businesses with at least five employees. The fieldwork ran from February to May 2025, so it's a snapshot of where UK firms were last year, but it is the largest official picture available.

How many UK businesses actually use AI?

Around one in six. The report found that 16% of businesses were using at least one AI technology, 5% planned to, and 80% neither used AI nor planned to.

Size makes a big difference:

For most of the SMEs that make up the UK's B2B market, AI adoption is still early. That's an opportunity: the firms that get it working well now are competing against a field that mostly hasn't started.

What are UK businesses using AI for?

Content and marketing lead the list. Among the 919 businesses that were using or planning to use AI, the report found the most common uses were:

• Creative and content creation: 77%
• Administrative and support tasks: 70%
• Data and analytics: 56%
• Automation: 47%

By business function, 72% said they use or will use AI in marketing, level with administration and ahead of IT at 64%. Note the base: these percentages are of businesses already using or planning to use AI, not of all UK firms.

The takeaway is simple. When a UK business adopts AI, marketing is usually first in line. It's also where mistakes are most public, because what marketing produces goes out under the company's name.

What's stopping businesses from using AI?

Not mainly money. Asked about barriers, 71% of all businesses said they had not identified a use for AI in their organisation, and 60% said they had limited AI skills, expertise or knowledge.

Among the smaller groups who raised specific concerns, those concerns ran deep. Of businesses that cited ethical concerns, 80% called them a significant barrier. The same was true for 76% of those citing cost and 72% of those citing unclear regulation.

In other words, most firms aren't held back by a budget. They're held back by not knowing what AI would do for them, and not having the skills to find out.

Why it matters that marketing is changing fast

The pressure is real even for firms that haven't adopted AI themselves, because their buyers have. HubSpot chief executive Yamini Rangan told investors in August 2025 that marketing is going through its "biggest shift in decades", as organic search traffic declines and more people turn to AI tools for answers. She said just 10% of HubSpot's leads now come from its blog.

So the question for a UK B2B business isn't really whether to use AI internally. It's whether your marketing still works when buyers research through AI answers rather than ten blue links. Our guide on how to get cited in AI search covers that side.

Where should a B2B team start?

The government data points to a practical answer: don't start with an AI strategy, start with one job. Based on the barriers above, this is the approach we'd take with a small B2B team:

1. Pick one repeatable task where a first draft is the slow part: case study write-ups, post-call follow-up emails, event recaps, LinkedIn posts from a longer piece.
2. Write a one-page AI policy. What tools are allowed, what data must never go into them (client details, anything confidential), and who checks the output. This answers most ethical and regulatory worries in plain terms.
3. Keep a person on every output. Google's own guidance now says to manually fact-check all AI-generated content before publishing, and the same discipline protects your brand everywhere else.
4. Measure time saved and quality, not just output. If it isn't saving hours or improving the work after a month, change the task.
5. Build skills as you go. The 60% skills gap closes fastest by using AI on real work with a clear brief, not through generic training.

If you'd like help choosing that first job and setting it up properly, get in touch.

Sources

• Department for Science, Innovation and Technology: AI Adoption Research, published 28 January 2026
• DSIT: AI Adoption Research Report (PDF), figures 1, 9, 13, 20 and 21
• CX Today: HubSpot has rebuilt its funnel, and it's working
• Google Search Central: Using generative AI content