An AI ad agency is an advertising agency that uses AI as part of how it makes, places or optimises ads. The term covers three quite different businesses: performance agencies that automate bidding, targeting and creative testing; AI production studios that generate film and imagery; and hybrid creative agencies, where human creative directors and AI agents work the same brief from insight to finished asset.
Brands searching for an “AI advertising agency” often get the first kind when they wanted the third. Here is how to tell them apart, what actually changes in the work, and what the UK and EU now expect.
What is an AI ad agency?
Every agency uses AI now. IAB research reports that 83% of ad executives say their company has deployed AI in the creative process, and in the UK ISBA found the share of advertisers with a live generative AI use case rose from 9% in April 2024 to 41% in July 2025. An AI ad agency is one where AI is not a side tool but part of how the business is organised, which changes what you buy, how it is priced and who is accountable.
What is AI advertising?
AI advertising is the use of artificial intelligence across the life of an ad: deciding who sees it, making it, and learning from how it performs. Two kinds of AI do that work. Predictive AI has run bidding, targeting and optimisation in programmatic media for more than a decade. Generative AI is the newer part: it writes copy and makes images, video, voice and versions, which is why it has changed the creative side of advertising rather than just the media side.
Adoption is no longer the question. In IAB's 2025 video ad spend report, 86% of buyers said they use or plan to use generative AI to build video ad creative, and they expect it to be behind 40% of all ads by 2026, with small and mid-sized brands moving fastest.
The better question is what makes it work, and the evidence points the same way: people and AI together, not either one alone. In a field experiment by Harang Ju and Sinan Aral at MIT, 2,234 people made 11,024 ads in human-human or human-AI teams. The human-AI teams produced 50% more ads per worker and stronger copy, which lifted click-through and view time when the ads ran on X. The human teams still made better images, and the more the work was handed to the AI, the more alike the ads became. Kantar's analysis of 356 AI-generated ads found the same spread of results as traditional advertising: consumer insight, bold creative and strong branding decided performance, not the production method.
So AI advertising at its best is a division of labour. AI widens the search, speeds up the making and multiplies the versions. People bring the insight, the taste and the final call.
How to use AI in ads: where to start
Start where AI is strongest and the risk is lowest, and keep a person accountable at every step.
- Start with versions, not the big idea. Cut-downs, formats, languages and market variants are where AI saves the most time with the least risk to the brand.
- Use AI to widen the search, then let people choose. AI can map a category and generate many more routes than a team would in the same time. Skill matters here: in a Journal of Advertising Research study of 98 advertising professionals, generative AI increased the number of ideas only for people who had been trained to prompt it.
- Know where AI is strong, and check where it is not. In a Harvard Business School and BCG experiment, consultants using GPT-4 completed 12.2% more tasks, 25.1% faster and at more than 40% higher quality on work inside the AI's capabilities, but were 19 percentage points less likely to get the right answer on a task outside them. A senior human editor is what catches the second case.
- Brief the AI like a team member. Clear, task-focused instructions work best: in the MIT experiment, task-oriented communication raised productivity and ad quality in human-AI teams without making the work look the same.
- Test in market, and decide disclosure at the brief. Run the variants, read the results and feed them into the next round. Settle how you will handle disclosure before production starts (see disclosure and regulation below).
The brands getting the most from AI are not replacing their creative teams; they are giving them a faster, wider way to work. That is how Brave New World® is built: human creative directors and eight AI agents on the same brief, in the Creative Intelligence System.
The three models
| Model | What it sells | Where AI sits | Where people sit | Best for |
|---|---|---|---|---|
| Performance automation | Media buying, targeting and ad testing | Bidding, audience modelling, automated creative variants | Strategy and account management | Brands whose main problem is efficiency of spend |
| AI production studio | Generated film, imagery and content at volume | The making itself | Prompting, editing and finishing | Brands that already have the idea and need it made fast |
| Hybrid creative agentcy | Strategy, ideas and production in one team | Research, ideation, scoring and production, through AI agents | The insight, the idea, the taste and every sign-off | Brands that need a better idea, not just more content |
Brave New World® is the third kind, and it is why we spell it agentcy: the agency runs on agents, and people lead it. The full definition is in What Is an AI Creative Agentcy?
What actually changes in the work
- Speed. Research and drafting that used to fill weeks now take hours, so a brief can move from insight to finished asset in days. At our studio, finished creative can be delivered in as little as two days.
- Versions. One idea can be adapted for every platform, format and market without a new production for each, which matters when one campaign runs across the UK and Europe.
- Testing. More ideas get explored and scored before anything is shown to a client. Ours are scored against a reference set of 3,000+ award-winning campaigns from 12 award shows.
- Cost structure. When production stops being the bottleneck, billing by the hour stops making sense. The honest model prices the outcome and the senior thinking, not the time around it.
What brands worry about
Brands are right to ask hard questions. In WFA research, eighty per cent of multinational brand owners said they had concerns about how their agencies use generative AI, fewer than half (44%) had policies or guidelines on its use, and 55% planned to review their agency contracts. The concerns are practical: who checked the work, which tools touched it, what data went in and who owns what comes out.
A good AI ad agency answers those questions before you ask them. If it cannot tell you where a human reviews the work, that is the answer.
How people react to AI ads
Audiences are more cautious than marketers. Kantar's Media Reactions 2025 found over 60% of consumers worry that generative AI could lead to fake or misleading advertisements, while over 70% of marketers embrace it for advertising creativity. In the UK, IPA research found 74% of consumers believe brands should be transparent in their use of AI-generated content. Trust is earned by quality and honesty, not by hiding the tools.
Disclosure and regulation for AI ads
This is a summary, not legal advice.
- UK, ASA and CAP. There is no blanket legal requirement in the UK to disclose AI use in ads. Ask two questions: is the audience likely to be misled without disclosure, and would a disclosure clarify or contradict the message?
- UK, industry principles. The Advertising Association's guide for responsible generative AI in advertising, and the earlier IPA and ISBA principles, put transparency and human oversight at the centre.
- EU, the AI Act. Article 50 requires deepfakes to be disclosed from 2 August 2026, with a lighter duty for evidently creative or artistic work. Providers of AI systems must also mark synthetic outputs in a machine-readable way; under the Digital Omnibus on AI, systems already on the market before that date have until 2 December 2026 to comply.
Frequently asked questions
What does an AI advertising agency do?
It uses AI as part of how it makes, places or optimises ads. Depending on the model, that means automating bidding and creative testing, generating film and imagery, or running human creative directors and AI agents on the same brief from insight to finished asset.
Is an AI ad agency cheaper than a traditional agency?
Usually, because AI removes much of the manual production and versioning work that used to be billed by the hour. The saving is real only if the agency prices to the outcome rather than to headcount, so ask how a quote is built.
Will AI ad agencies replace traditional agencies?
They are replacing the traditional operating model rather than the need for agencies. Big teams and billable hours are giving way to smaller senior teams working with AI agents. Brands still need people who can decide what to say and judge whether the work is good.
Do AI-generated ads need a label in the UK?
Not automatically. The ASA and CAP say there is no blanket legal requirement in the UK to disclose AI use in ads; the question is whether the audience would be misled without disclosure. Campaigns that also run in the EU must meet Article 50 of the EU AI Act from 2 August 2026.
Who owns the work an AI ad agency makes?
Ownership of client deliverables should be written into the statement of work for each project, including which tools were used and how. At Brave New World, the Creative Intelligence System and its agents are the intellectual property of Brave London Ltd, and deliverable ownership is set out per project.
Further reading: How to Choose an AI Creative Agency and AI in Creativity: What It Changes and What It Can't. Have a brief? Send it to us: we come back within one working day.