AI in Advertising Agencies: What’s Changing

AI in advertising agencies is changing how work gets done, but not always in the ways the vendor decks suggest. The shift is less about replacing creative talent and more about removing the low-value labour that surrounds it: briefing admin, asset versioning, performance reporting, audience segmentation, copy iteration. Agencies that understand this distinction are moving faster. Those treating AI as a positioning statement are mostly just talking about it.

This article covers where AI is creating genuine operational change inside agencies, where the hype still outpaces the reality, and what senior marketers should be asking before they commit budget or restructure teams around it.

What Is AI Actually Doing Inside Agencies Right Now?

Strip away the conference presentations and the LinkedIn announcements, and the honest picture of AI inside most agencies in 2025 is fairly unglamorous. It is being used to draft copy faster, generate image concepts for client review, summarise research documents, build first-pass performance reports, and produce multiple ad variations for testing. None of that is revolutionary. All of it is useful.

The agencies getting the most out of these tools are not the ones with the biggest AI budgets. They are the ones that identified specific friction points in their workflows and applied tools to those points deliberately. That is a boring answer, but it is the right one.

When I was running iProspect, a significant portion of the team’s time was consumed by work that required intelligence but not creativity: pulling data from multiple platforms, formatting it into reports, writing the same client-facing narrative with different numbers each week. That work was expensive in human hours and unrewarding for the people doing it. AI tools applied to that specific problem would have freed up real capacity. Not headcount reduction, capacity reallocation toward the strategic and creative work that actually builds client relationships and drives retention.

That is the framing worth holding onto. AI in agencies is a capacity question before it is a capability question.

If you want a broader view of how AI is reshaping marketing operations beyond the agency context, the AI Marketing hub at The Marketing Juice covers the strategic and practical dimensions across the full marketing mix.

Where Are Agencies Finding Real Efficiency Gains?

The efficiency gains showing up consistently across agencies fall into a few clear categories.

Content production at scale. Agencies running paid social or performance campaigns are using AI to generate dozens of copy variants from a single brief. Testing velocity has increased meaningfully for teams that have set this up properly. The caveat is that more variants require more structured testing frameworks. Generating fifty versions of an ad and running them without a clear hypothesis is not efficiency, it is noise.

Research and briefing compression. Desk research that used to take a junior strategist half a day can be compressed significantly with the right AI tools. Competitive landscape summaries, category trend overviews, audience behaviour synthesis. The output still needs human review and editorial judgment, but the starting point is better and faster. Moz has written thoughtfully about where AI content creation adds value and where it falls short, and the principle applies equally to research as to copy.

Performance reporting. Automated report generation has been around for years, but the quality of narrative commentary that AI can now produce around data sets has improved substantially. Agencies are using this to reduce the time their analysts spend writing the same paragraph every month and redirecting that time toward interpretation and recommendation.

Asset versioning and localisation. For agencies working across markets or producing campaigns that need to run across multiple formats, AI-assisted versioning has reduced the manual production burden. This is particularly visible in agencies with large retail or FMCG clients where the volume of executions is high and the creative variation is relatively constrained.

SEO and content strategy support. Tools that help agencies identify content gaps, cluster keywords, and model search intent have become standard in most digital agencies. Ahrefs has produced useful material on how AI tools are being applied to SEO workflows that gives a clear picture of where the practical value sits.

Is AI Changing the Agency Business Model?

This is the question that agency leaders are not always comfortable answering honestly in public. The business model tension is real.

Most agencies have historically charged for time. The more hours a project requires, the more revenue it generates. AI compresses time. If a campaign that used to take forty hours of production can now be delivered in twenty, the agency has two choices: charge less, or use the freed capacity to do more or better work for the same fee.

The agencies that are handling this well are the ones shifting their value proposition away from execution volume and toward strategic and creative judgment. They are positioning the time savings as a benefit to the client, not a threat to their margin, because they are reinvesting that capacity into the work that genuinely requires human expertise.

The agencies that are struggling are the ones whose value was primarily executional. If your agency’s main offer was producing a lot of content quickly at a competitive rate, AI has put pressure on that model in a way that is not going away. Clients can now access tools that produce competent executional work without agency involvement. The question those agencies need to answer is what they offer that the tool cannot.

I have seen this pattern before with other technology shifts. When programmatic advertising matured, the agencies whose value was in manual media buying were exposed. The ones that survived and grew were the ones that moved their expertise up the value chain into strategy, audience intelligence, and measurement. The same logic applies here.

What Are Agencies Getting Wrong About AI Adoption?

There are a few consistent mistakes I see when agencies talk about their AI adoption.

Treating AI as a marketing message before it is an operational reality. I have read agency positioning decks that lead with AI-powered everything before the agency has actually embedded a single tool into a repeatable workflow. Clients are not impressed by the claim. They are impressed by faster turnaround, better output, and sharper thinking. If AI is helping you deliver those things, say so. If it is not yet, stop leading with it.

This reminds me of a pattern I saw repeatedly when judging the Effie Awards. Agencies would submit campaigns built around innovation for its own sake. VR installations, experimental formats, technology-first executions that had no clear connection to the business problem the client actually needed to solve. The question I kept asking in the judging room was: what problem is this solving? AI adoption in agencies risks the same trap. The technology is interesting. The business case has to come first.

Adopting tools without redesigning the workflow around them. Dropping an AI writing tool into an existing content process and expecting the team to figure it out is not a strategy. The tool needs to sit inside a redesigned process that accounts for where human judgment is required and where it is not. Without that, you get inconsistent output and frustrated teams.

Underestimating the quality control requirement. AI output needs editing. Not light editing. Proper editorial review by someone who understands the brand, the audience, and the commercial objective. Agencies that have skipped this step have produced work that is technically competent and strategically empty. Semrush has covered this tension well in their writing on how to build content strategies around AI optimisation tools, and the core point holds: the tool supports the strategy, it does not replace it.

Ignoring the talent implications. Some agencies have used AI adoption as cover for reducing junior headcount. That is a short-term financial decision with a long-term talent pipeline problem. Junior roles are where senior talent develops. If you remove the entry-level executional work without replacing it with structured learning opportunities, you hollow out your future capability. The agencies that are handling this thoughtfully are redesigning junior roles rather than eliminating them.

How Should Agencies Be Thinking About AI and Creative Work?

Creative work is where the AI conversation gets most heated and least useful. The debate tends to polarise around two positions: AI will replace creatives, or AI is just a tool and creatives have nothing to worry about. Neither is quite right.

The honest position is that AI is already changing what creative roles look like, and it will continue to do so. The question is not whether creative work will change but which parts of it will change and which will not.

AI is reasonably good at generating options within a defined creative territory. Give it a clear brief, a brand voice, a format constraint, and it can produce a range of executions quickly. That is useful for production, for testing, for filling a content calendar. It is not useful for defining the creative territory in the first place. That requires the kind of strategic and cultural intuition that comes from understanding people, not pattern-matching text.

The creative directors I respect are not worried about AI replacing their thinking. They are using it to remove the friction between thinking and making, which means they can test ideas faster and spend more time on the ideas worth testing. That is a genuine improvement in how creative work gets done.

HubSpot has done useful work cataloguing how generative AI video tools are being applied in marketing production, and the pattern is consistent: the technology is accelerating production, not replacing the creative judgment that determines what gets produced.

What Do Clients Actually Want From Agencies Using AI?

Clients want two things, and they have always wanted these two things: better work and better value. AI is relevant only insofar as it delivers one or both.

Most clients are not asking their agencies to explain their AI stack. They are asking whether the brief was understood, whether the work is on strategy, whether the campaign performed, and whether the agency is a partner worth keeping. AI is infrastructure, not the offer.

Where clients are starting to push back is on fees for work that AI has made faster to produce. If a content agency was charging for twenty hours of copy production and AI has reduced that to eight hours, clients will eventually notice and ask questions. Agencies that are ahead of this conversation are having it proactively, reframing the value around the quality of thinking rather than the volume of hours.

There is also a transparency question that is becoming more prominent. Some clients want to know when AI has been used in the production of their work. Agencies need a clear policy on this, not because clients are universally opposed to AI involvement, but because the relationship depends on trust and trust depends on transparency.

Early in my career, I learned that the fastest way to lose a client relationship was to let them discover something important that you had not told them. The same principle applies here. If AI is part of how you are delivering their work, that is a conversation to have, not a fact to obscure.

What Does the Next Phase of AI Adoption Look Like for Agencies?

The current phase of AI adoption in agencies is largely about tool integration at the task level. Individual team members using AI to do specific tasks faster. The next phase, which is already beginning in more operationally mature agencies, is process-level integration. AI embedded into the workflow at multiple points, with human checkpoints designed into the process rather than bolted on as an afterthought.

Beyond that, the agencies thinking furthest ahead are looking at how AI changes the nature of the client relationship itself. If AI can handle a significant portion of the executional work, what does the agency become? The answer, for the best agencies, is a strategic and creative partner that is less dependent on production volume for its revenue and more focused on the judgment-intensive work that genuinely requires experienced people.

Semrush has written about where AI optimisation tools are heading in terms of capability development, and the trajectory points toward more autonomous execution of defined tasks. That raises the stakes for agencies to be clear about what they are offering that sits above the line of what AI can automate.

The agencies that will be in the strongest position in five years are the ones that have used this period to get sharper about their strategic and creative value, not the ones that have used AI primarily as a cost-reduction mechanism. Cost reduction is a defensive move. Sharpening your value proposition is an offensive one.

Moz has published a useful Whiteboard Friday on generative AI for SEO and content success that is worth reviewing if you want a grounded view of where the content and search dimensions of this are heading. The framing there, that AI accelerates the production of content but does not replace the strategic thinking behind it, is consistent with what the best agencies are finding in practice.

There is more on the strategic and commercial dimensions of AI in marketing across the full AI Marketing section of The Marketing Juice, covering everything from workflow design to tool selection to the measurement questions that most vendors prefer not to answer.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions