AI Search
The traditional agency model is set to change as AI-powered agencies opt for a leaner, more creative model.
If you believe the headlines, nobody is using AI, the tech is overhyped, and adoption has stalled. But if you talk to actual marketers, you hear something else entirely.
Quiet confessions about doubling their output. Creatives who now storyboard with AI. Strategists who get to a decent first draft in minutes rather than days. Productivity is up, is the headline news, even if nobody is brave enough to put the real numbers in a board pack yet.
Let’s use myself as an example. I am writing this using voice typing. Not because it is futuristic, but because it is faster than my two-finger keyboard technique.
Years ago, when I was a police officer, I asked to go on a typing course. The answer was no. The productivity gain could not be “monitored or measured”, so it was not worth the money. We had a typing pool and therefore officers did not need to type.
A few years later, every ‘young’ new recruit could touch type. Then everything moved to mobiles, and those same officers now file statements with their thumbs at a speed my old inspector would not have believed possible.
The work changed. The interface changed. The job did not.
We still needed judgment, experience and thinking. We just finally stopped wasting time sitting at a desk, with a keyboard, punching keys.
The thumb and the small screen, combined with predictive text (plus years of training through social media posts), have allowed the next generation to increase productivity.
That is exactly where agencies are heading with AI. The tools will quietly transform the doing. The questions are which agencies are ready to rebuild themselves around the thinking and what will work look like for those agencies?
We need to be clear about the enemy. AI is not the enemy here. The enemy is the traditional agency model that surrounds it.
The big glass building in the city. The channel silo structure. The headcount status symbol. The costly signal of an expensively designed reception that tells clients, “We are safe, we have enough clients to pay for this.”
That model made sense when work was manual. You needed a room full of people to plan, build and traffic a campaign. Today, AI can do all of that. It can:
● Write decent blog drafts in 30 minutes
● Generate product images, variants and headlines on demand
● Build and optimise a paid campaign with almost no human clicking.
So the advantage moves. It shifts from owning bodies and channels to owning brains and ideas and the lean AI agency starts from that premise.
They are not built around “our SEO team”, “our social team”, “our performance team”. They are built around a small group of marketers who understand segmentation, targeting, positioning and marketing effectiveness, and, in turn, surrounding those people with AI agents, tools and workflows.
They exist to think. The machines exist to do.
If AI does most of the typing, why are so many agencies still paying for buildings designed as typing factories?
In a lean AI agency, the big office is the first thing to go. Not because offices are bad, but because most of what happens in them is now pointless. There is no logic in commuting across a city to sit in silence, wearing headphones, speaking into AI tools you could access from home.
The office of a lean agency is different. It is small, branded and built for one job: human-to-human collaboration.
The new marketing buzzword is GEO – but is it real?Think of a single studio space. Sofas, whiteboards, a huge digital wall, a few tables, coffee, maybe some playful touches. Clients are invited not for “updates”, but for working sessions. Everyone talks, sketches, argues, prioritises. AI records everything, tidies it, drafts the follow-ups and executes the grunt work later.
You are not paying for square footage. Instead, you are paying for the ideas that come out of the room. In other words, you are paying for the brilliance and tacit knowledge of experts.
The costly signal of a big building is replaced with a better one. A tight, distinctive, creative space that says this is a place where thinking happens.
For twenty years, agencies have sold specialism. Facebook ad specialists, SEO specialists, programmatic specialists. Each with their own process, acronym and slide template.
The problem is that most of the “secret knowledge” inside those disciplines now lives in AI prompts and public documentation. If you want to know how to structure a Google Ads campaign, you can ask a model. If you want to know the standard SEO playbook, it will summarise the best blog posts for you.
That does not make practitioners irrelevant. Rather it makes mechanics a weak source of advantage. In a lean AI agency, the value sits with people who can:
● Diagnose the market and the buyer
● Segment and choose a target properly
● Position a brand in a way that is actually distinctive
● Decide the right balance between brand, activation and channels.
Juniors still exist, but as apprentices, learning judgment and marketing, not just where to click in an ad manager.
Agencies can implement structured upskilling programs for their existing staff to navigate this transformation. Training sessions on leveraging AI tools efficiently, as well as workshops focused on developing strategic thinking and creative problem-solving, can be part of this initiative.
By equipping their teams with these skills, agencies can create a seamless transition where talent thrives alongside technology. Specialism does not disappear. It just stops being the front of house. The generalist strategist sits at the centre, AI multiplies their impact across every discipline.
As AI accelerates content production, attention becomes the constraint. Every category is about to be buried under competent, personalised, instantly generated material.
That is why the lean AI agency of the future looks suspiciously like a PR shop with ad capabilities. Its goal is not to fill the content calendar. Its goal is to create fame. To generate ideas that earn coverage, get talked about, and are worth watching and sharing.
AI will cut the edit, version the idea for every platform, transcribe the interviews and repurpose the footage. The hard bit is the spark.
Look at what has happened on the high street. The old kebab shop has been reborn as the smash burger joint. Same basic inputs, but suddenly there is neon, spectacle, indulgent milkshakes, dirty fries and queues out of the door. The product is not just food, it is an experience.
How to engineer fame in an AI era – and why you need to start nowLean AI agencies will do the same to themselves. Highly distinctive brands with strong visual identities and clear points of view. Places clients actively want to spend time in because that is where the interesting ideas come from.
In that world, a lean AI agency is not selling “PPC management” or “SEO packages” instead they are selling the ability to create work that gets noticed and moves the dial – they can then use AI to deliver it faster and more profitably than any traditional rival.
Will this be good for jobs? Probably not in the short term.
A lean AI agency can achieve the output of a traditional shop with a fraction of the headcount. That is precisely why it can undercut, out-think and outlast them. But effectiveness still rules.
As Les Binet likes to remind us, effectiveness comes first, and efficiency second.
The point of AI is not to make bad work cheaper. It is to liberate talent from the keyboard so they can spend more time on diagnosis, strategy and big ideas.
Agencies that cling to big offices, channel silos and bloated structures will discover the cost side of AI before they feel the benefit. The ones that rebuild themselves as small hubs of serious marketers, amplified by AI, will set the pace.
The more artificial intelligence reshapes execution, the more human intelligence will decide who wins. The question isn’t if this will happen. The real question is how quickly.
Marketing Week’s weekly round-up of the technology stories that impact the marketing sector: from AI to martech, regulation to public perceptions.
With Google’s AI Mode now live in the UK, experts suggest brands are “overwhelmingly underprepared”.
In a volatile world where shareholders demand quick wins and squeezed consumers want short-term savings, how can marketers champion long-term thinking?
From developing a “common playbook” for success to establishing a sense of shared context, B2B sales and marketing can become close collaborators.
The grocer is slashing prices on “thousands” of everyday products in a bid to undercut loyalty card prices offered by rivals after a difficult December.
Data finds over half of ethnic minority workers have been paid less than a colleague from a different ethnic background for similar work.
Primark wants to cement its status as a “value leader” after a “challenging start” to the year, with 2026 profits expected to dip below last year’s levels.
