Writing

• AI in Marketing
• 30 September 2026

How to Use AI in Marketing Without Buying 40 Tools You Never Open

To use AI in marketing well, start with the tools your organisation already pays for, pick two or three recurring pieces of work where AI can take on a clear, checkable step, write down who reviews what, train the team properly and measure the result against a baseline for 90 days. Only then decide what to buy. Most teams do this in reverse.

Gartner’s 2026 CMO Spend Survey of 401 marketing leaders in North America, the UK and Europe found that AI now takes 15.3% of marketing budgets, yet only 30% of CMOs report mature AI readiness capabilities. So most of us are paying for AI we aren’t set up to use, which is a very expensive way to own a login.

Why do most attempts to use AI in marketing stall?

Usually because the starting move is a purchase rather than a decision about work. Someone sees a demo, a free trial becomes a subscription, another team buys something that does roughly the same thing, and a year later the finance team finds nine AI writing tools on the card statement and nobody can say which campaign any of them helped. We then blame the tool, the vendor, the data, the agency, probably the weather.

The Office for National Statistics’ July 2026 analysis of its Business Insights and Conditions Survey (38,637 responding businesses) puts AI use at around 35% of UK firms with 10 or more employees, up from around 12% in late 2023. But the average adopting business uses 1.6 AI technologies, only 10% describe their use as extensive, and only 11% have trained more than half their workforce. That is a lot of people with access and not many with a habit.

Marketing Week’s report on the Chartered Institute of Marketing’s research with YouGov (500 marketing decision makers) found that 47.2% admit to using AI without a defined skills strategy, and 66.5% of the 2,350 respondents to Marketing Week’s 2026 Career and Salary Survey have noticed an AI skills gap in their team. Buying more software doesn’t close either of those gaps.

Gartner’s survey of 413 martech leaders, run between June and August 2025, found that 81% were piloting or had implemented AI agents, and 45% of those said the vendor-offered agent capabilities did not meet their expectations. Half said they lacked the technical infrastructure and data foundation to deploy agents properly. What an agent actually does differently from a chat assistant is covered in generative vs agentic AI in marketing.

How can AI be used in marketing, realistically, today?

If you want the definitions first, the pillar piece on what AI in marketing actually is covers them. For a starting plan, the more useful question is which jobs AI does reliably enough to hand over a step, and which ones it does convincingly enough to fool you. Gartner’s martech survey gives a sense of where leaders are pointing agents first: 52% named content or marketing asset production as a top use case, 49% content enrichment and 43% campaign management and optimisation.

The table below is how I’d sort the common uses for a team starting out. The right-hand column matters most, because it’s where the time saving either survives or gets eaten by rework.

Adam Peruta and Carrie Riby of Syracuse University’s Newhouse School worked with Ipsos on a study of 20 video ads from 10 brands: 10 original human-made ads and 10 AI versions built by reverse-engineering the originals’ briefs with Google’s Gemini and generating the new ads with OpenAI’s Sora 2. Tested with 3,000 US consumers, the human-made ads came out 14% stronger on short-term effectiveness and 17% stronger on long-term brand effect, according to Ipsos’s write-up. Only 25% of people who saw an AI ad were even somewhat confident it was AI-made. Peruta summarised the work in HBR in September 2026, arguing that AI creative is a governance issue as much as a production tool. What worries me about that result is the timing: the short-term gap shows up in a pre-test if you run one, but the brand equity gap is the kind you find out about a year later when the tracker dips and we’ve already blamed the media plan.

How to use AI in marketing: a 90-day starting plan

This is the sequence I’d use for a marketing team of three to fifty people. It assumes no dedicated AI budget, which is the realistic position for most UK teams outside the large enterprises Gartner surveys.

1. Audit what you already pay for (week 1)

If your Microsoft 365 licences include Microsoft Copilot Chat, Microsoft’s enterprise data protection documentation states that prompts, responses and data accessed through Microsoft Graph aren’t used to train foundation models. If your Google Workspace plan includes Gemini, Google’s Workspace privacy hub says your content isn’t human reviewed or used for model training outside your domain without permission. If the team prefers ChatGPT, OpenAI’s enterprise privacy commitment that it does not train on your data by default covers the business tiers (ChatGPT Business and Enterprise, among others) and not someone’s personal account, which is the whole reason to standardise.

Then ask people, without judgement, which tools they’re already using on their own cards and for what. That list tells you more about real demand than any vendor roadmap.

2. Map the recurring work (weeks 1-2)

Get each person to list the five recurring tasks that eat the most time each month. Score each one on three things: hours per month, how easy it is to check the output, and what it costs if the output is wrong. The sweet spot is high hours, easy to check, low cost of error. Monthly reporting commentary, turning one approved asset into channel variants and summarising customer feedback almost always score well. Brand positioning, pricing and anything that makes a regulated claim almost always score badly, and that’s fine: they stay human.

3. Pick two or three workflows and redesign them (weeks 2-4)

The temptation is to bolt AI onto the existing process and call it done. McKinsey’s State of AI 2026 survey of 1,719 respondents found that 73% of its “high performers” (the 6% attributing 5% or more of EBIT to AI) are fundamentally redesigning workflows, against a quarter of everyone else. Masha Shunko (University of Washington) and Serguei Netessine (Wharton) make the same case in HBR’s “Stop Automating Old Processes. Design New Ones Instead.” and give a handy way to decide the human role at each step: assist, approve, audit or automate.

On the monthly performance report, AI can be allowed to automate the data pull and assist with the first commentary draft, while an analyst approves the narrative before it goes to the leadership team. On social variants, AI drafts, the channel owner approves, and someone audits a sample each month for tone drift. Writing those four words next to every step sounds bureaucratic and takes about twenty minutes. I’ve set out a fuller method in how to build an AI marketing workflow.

4. Write the one-page rules (week 3)

Keep it to a page or nobody will read it. Cover which tools are approved and on which accounts, what data never goes into a prompt (customer personal data, unreleased financials, anything under NDA), who checks factual and legal claims, and when you disclose AI use. On that last point, the ASA and CAP’s May 2025 guidance is clear that there is no blanket legal requirement in the UK to disclose AI in ads, that the CAP Code applies the same way whether AI was used or not, and that the test is whether the audience is likely to be misled without disclosure. On personal data, UK GDPR applies as it always did; the ICO’s guidance on AI and data protection is the reference, though the ICO notes it is under review following the Data (Use and Access) Act, so check it again before you finalise.

5. Train people on your own briefs (weeks 3-6)

BCG’s June 2026 AI at Work survey of close to 12,000 employees found that only 36% feel they’ve received adequate upskilling, and 66% get limited or no guidance on what to do with the time they save. That second number is the one I’d fix first. A good session takes one of your chosen workflows, runs it live with the real brief and the real data, and ends with everyone having a saved prompt they’ll use the following Monday. Two I’d start with:

You are helping a UK marketing team write its monthly performance summary.
Here is the channel data for [month] and the same month last year: [paste].
Here are the three objectives for this quarter: [paste].
Write a 200-word summary for the leadership team in plain British English.
Separate what changed from why it might have changed.
For every "why", say what evidence in the data supports it, or write
"not supported by this data". Do not invent figures.

Here are 150 customer comments from our post-purchase survey: [paste].
Group them into no more than six themes. For each theme give the number
of comments, two verbatim quotes, and one thing it suggests we test.
List separately any comment that doesn't fit a theme but could matter
(complaints about safety, legal, accessibility or pricing errors).

The last line of the second prompt is there because AI summaries are very good at making the awkward comment disappear, and the awkward comment is often the one you needed.

6. Measure against a baseline, then decide what to buy (weeks 4-12)

Record how long each chosen workflow took before you changed it, and its rework rate. Re-measure at 30, 60 and 90 days. Hours saved is the easy number; the better ones are rework rate, turnaround time and whether the output performs as well as the old version. The method for putting a financial value on that is in measuring the ROI of AI in marketing. Only at day 90, with that evidence, does it make sense to pay for a specialist tool, and at that point you’ll be buying it for a named workflow with a known saving, which is a far stronger position to negotiate from than “the demo looked great”.

Where does AI fit in your marketing strategy?

AI in marketing strategy has two separate meanings, and mixing them up causes most of the confusion. One is using AI to help do strategy: faster research, competitor summaries, scenario drafts, stress-testing a positioning statement. The other is deciding what role AI plays in how your marketing function operates, what it costs and what it’s for. Both belong in the plan, but they need different owners.

On the first, AI makes a decent research assistant. It can summarise forty competitor websites before lunch, but it can’t tell you which of your three possible positionings your sales team will actually sell, because that depends on things that aren’t written down anywhere. My view, from pharma brand management at Cipla and Alembic and now leading AI adoption in marketing at ACTAsia, is that the choice stays with people who’ll be accountable for it, and AI gets the inputs ready.

On the second, the strategic question is where AI changes your cost base or your growth, with a number attached to each. Marketing leaders Gartner surveyed between August and October 2025 (402 CMOs) expect the share of marketing work automated by AI to rise from 16% in 2026 to 36% by 2028, yet 80% said staff fear and anxiety were a barrier to experimentation. And in the 2026 spend survey, organisations with mature AI capabilities put 21.3% of marketing budget into AI against the 15.3% average. McKinsey’s survey adds that respondents most commonly report revenue gains from AI use in marketing and sales. If you want the wider numbers in one place, they’re in AI in marketing statistics 2026.

In practice I’d put three lines in the marketing plan: the two or three workflows being redesigned this year with their target saving, the one strategic question AI research will support (a segmentation refresh, say), and the rules and training commitment. It’s not glamorous, and it won’t get a round of applause at the offsite, but it will still exist in March.

Getting started: a first-month checklist

• List every AI tool the team uses, paid or free, work or personal account.
• Confirm which approved tools your existing Microsoft, Google or OpenAI licences already cover, and on what data terms.
• Collect each person’s five most time-consuming recurring tasks, with hours per month.
• Score them on hours, ease of checking and cost of error; choose two or three.
• For each chosen workflow, mark every step as assist, approve, audit or automate, and name the approver.
• Time the current version of each workflow and note the rework rate.
• Write the one-page rules: approved tools, banned data, claims checking, disclosure.
• Run one live training session per workflow using real briefs, and save the prompts somewhere shared.
• Book the 30, 60 and 90-day reviews in the calendar now, before enthusiasm drops.
• Put a freeze on new AI subscriptions until the 90-day review, with a named person who can grant exceptions.

The freeze is the item people push back on, and it’s the one that stops you arriving at month twelve with forty subscriptions and a shared password spreadsheet 🙂

FAQs

What is the first step to using AI in marketing?

Audit what you already have. Organisations on Microsoft 365 or Google Workspace may already have an AI assistant available under business data terms, and staff may be using personal tools on their own. Listing both shows real demand and data risk before you spend anything. Then pick two or three recurring tasks that are time-consuming, easy to check and low-risk if wrong.

How much should a marketing team spend on AI to start?

For a first 90 days, often nothing beyond existing licences. Gartner’s 2026 CMO Spend Survey found AI takes 15.3% of marketing budgets on average, but only 30% of CMOs reported mature AI readiness. Proving value on two or three workflows with tools you already pay for gives you the evidence to justify specialist spend later.

Can AI replace a marketing agency or creative team?

Not for finished creative, on current evidence. In a Syracuse University and Ipsos study of 20 video ads tested with 3,000 US consumers, human-made ads were 14% stronger on short-term effectiveness and 17% stronger on long-term brand effect than AI versions built from the same briefs. AI is more useful for concepts, variants, research and production support.

Do UK marketers have to disclose when an ad uses AI?

There is no blanket legal requirement in the UK to disclose AI use in ads, according to ASA and CAP guidance published in May 2025. The CAP Code applies the same way whether or not AI was used, and disclosure is needed where the audience would otherwise be misled. Disclosure does not fix an ad that is misleading in the first place.

How long does it take to see results from AI in marketing?

Time savings on well-chosen workflows usually show within the first month, which is why a baseline matters. Quality and performance effects need longer, so review at 30, 60 and 90 days. BCG’s 2026 AI at Work survey found 42% of frontline employees who use AI regularly save eight hours a week, but also that 66% of employees get little guidance on using saved time.

Sources

1. Gartner, Gartner 2026 CMO Spend Survey Finds CMOs Allocate 15.3% of Marketing Budgets to AI, but Only 30% Are Ready to Scale AI Capabilities, 11 May 2026.
2. Gartner, Gartner Survey Reveals Marketing Leaders Expect AI Automation of Marketing Work to Double to 36% by 2028, 11 May 2026.
3. Gartner, Gartner Survey Finds 45% of Martech Leaders Say Existing Vendor-Offered AI Agents Fail to Meet Their Expectations of Promised Business Performance, 29 October 2025.
4. McKinsey & Company, The state of AI in 2026: On the road to ROI, 25 August 2026.
5. Office for National Statistics, Artificial intelligence in UK businesses: 2023 to 2026, 20 July 2026.
6. Ipsos, AI ads are good enough, and that’s the problem, 18 May 2026.
7. Harvard Business Review, Adam Peruta, Research: AI-Generated Ads Perform Worse Than Human-Made Ones, Even When Customers Can’t Tell Them Apart, 3 September 2026.
8. Harvard Business Review, Masha Shunko and Serguei Netessine, Stop Automating Old Processes. Design New Ones Instead., 14 September 2026.
9. BCG, AI at Work: Why Strategy Matters More Than Tools, 2 June 2026.
10. Marketing Week, Molly Innes, Just 5% of marketers expect AI to create new roles or opportunities, 9 June 2026.
11. ASA and CAP, Disclosure of AI in advertising: striking the balance between creativity and responsibility, 29 May 2025.
12. Information Commissioner’s Office, Guidance on AI and data protection, last updated 15 March 2023 (under review following the Data (Use and Access) Act).
13. Microsoft Learn, Enterprise data protection in Microsoft Copilot and Microsoft Copilot Chat, updated 18 August 2026.
14. Google Workspace Help, Generative AI in Google Workspace Privacy Hub, accessed 30 September 2026.
15. OpenAI, Enterprise privacy at OpenAI, updated 8 January 2026.