Foundations
Advertising creative tools demand signals: four dated UK official records and what each can and cannot show
Four dated UK records on AI use, the creative industries and online audiences, ranked by closeness to marketing work, with what each can show about tool demand.
The strongest public signal is business use of AI in marketing, followed by the scale of the creative industries, the online audience and earlier survey evidence. None proves demand for a particular product in England. Together they show where audience research deserves attention.
What to take away
- The four records show where to research demand, not proof of demand for any product.
- Marketing is the most common area of business AI use in the 2025 DSIT survey, at 72% of businesses using AI.
- Most AI use is text generation, so image and video tool demand needs separate evidence.
- Older survey evidence is kept for comparison, never added to newer figures.
How the four records were selected
Inclusion required a live primary record from the public body that produced the measure, a stated reference period and a link to marketing work, creative output or online audiences. Official publications were preferred because their method is published.
The ranking reflects closeness to marketing work, not market size. Excluded were vendor surveys, search-volume tools, commentary, undated summaries and any figure that could not keep its original geography. No England-only series was found, so each record covers the whole UK.
| Rank | Record | Reference period | What it can show |
|---|---|---|---|
| 1 | DSIT AI Adoption Research | Fieldwork February to May 2025 | Share of businesses using AI, and for what |
| 2 | ONS AI uptake and sentiment | April 2023 survey | An earlier baseline |
| 3 | Creative industries sector vision | Figures for 2021 | Scale of the sector |
| 4 | Ofcom Online Nation | Report updated December 2025 | Where to find audience data |
1. Business use of AI in marketing
The DSIT AI adoption research reports that around 1 in 6 businesses (16%) currently use at least one AI technology. Among those, marketing is the most common area of use, at 72%, level with administration. IT follows at 64%.
The fieldwork ran by telephone from 12 February to 2 May 2025, and 3,500 interviews were completed with businesses. That is a published method with a stated period, which makes the record easy to date and re-check.
Most use is text-based. Among businesses using AI, 85% use natural language processing and text generation. That points to copy and script work, not to image or video tools. A creative software seller should therefore treat the figure as a prompt for interviews about drafting, not as proof of demand for asset production.
2. An earlier baseline for comparison
The ONS analysis of AI uptake of its business survey, collected between 3 and 16 April 2023, found that 16% of UK businesses were using at least one AI technology asked about.
The share matches the 2025 figure, but the surveys differ in sample, question wording and method. Do not read the match as a trend line. Use the 2023 record to see how fast the question changed, and keep each figure with its own date and publisher.
3. Scale of the creative industries
In 2021 the creative industries generated £108 billion in economic value and employed 2.3 million people, up 49% since 2011. Advertising, marketing and creative digital innovation support sectors across the economy.
These are 2021 figures, so they describe the base, not current demand. They say nothing about software buying. Their use is context: they show a large workforce whose tools and workflows are a fair subject for customer interviews.
4. The online audience
The annual Online Nation report covers what people do online, how online providers serve them and their attitudes to the internet. The overview page dates from December 2025 and leads to the full report.
Audience time is not buyer demand. The report describes people across the UK, not software buyers, and it carries no view on campaign goals. Use it to pick formats and channels to ask about, then check the AI applications that suit each one.
Turn signals into a buying test
Choose one buyer group and one recurring task suggested by the evidence. Ask for a recent example, its asset count, revision path and approval owner. Then compare tools against a written selection checklist rather than a demo.
Seek a paid pilot with a stop rule agreed in advance, as set out in the ninety day plan. Public records point to sensible places to look. Only customer evidence can show that the problem is frequent, funded and poorly served.
Repeat the search when a publisher changes its method or issues a new reference period. Keep old and new definitions with the research log, and check the wider trends and outlook for context, so a change in wording is not mistaken for buyer behaviour.
Before you act
- Choose one buyer group and one recurring task from the evidence.
- Ask for a recent example, asset count, revision path and approval owner.
- Seek a paid pilot with a stop rule agreed in advance.
- Repeat the search when a publisher changes its method or period.
- Keep old and new definitions with the research log.
Common questions
Why rank AI use in marketing first?
It is the record closest to marketing work. The DSIT survey measures use by businesses, with a stated method and period. It still says nothing about creative software purchases or England alone.
Can the 2023 and 2025 figures be compared as a trend?
Not safely. Both show 16%, but the surveys use different samples and questions. Keep them as separate dated records.
What do the creative industries figures prove?
They show scale in 2021, not demand. Use them as context for interview questions about workflow and tools.
