The trillion dollar question about AI is addressed by Sir Martin Sorrell, founder and executive chairman of S4 Capital. He outlines five areas where AI is already reshaping agencies. First is visualisation and copywriting. So, if Monks is making an ad, instead of it costing USD 2.5-million and taking four months in Eastern Europe, the agency can now do it for USD 500,000 or even less, in four weeks. He says one packaged goods client averaged 200 days to produce an ad. But Monks produced an ad for them within 12 days.
As time shrinks, time-based billing erodes. "Our revenue is going to be less unless we shift the model to an output-based model, which is producing a million assets at a cost per asset," Sorrell stated.
Second is talent composition with fewer art directors and copywriters. But personalisation at scale, or what he calls "Netflix on steroids" — using first-party data and platform signals — could increase employment in data and tech.
Third is media investment management. The industry employs 250,000 people in planning and buying. Sorrell pointed out that BlackRock manages USD 14-trillion on its own in the investment industry, doing things algorithmically. The analogy is deliberate: capital allocation via better algorithms.
Fourth is operational efficiency. An outside broadcasting initiative with Nvidia, Amazon Web Services, and Adobe aims to cut costs by 80–90%. AI is less about novelty, he argues, and more about workflow simplification.
Fifth — and most disruptive — is organisational design. "Nvidia CEO Jensen Huang has around 50 to 60 executives directly reporting to him. Classic McKinsey management would have a maximum of 12. AI enables you to enfranchise people if you give them access to all the information," Sorrell stated. Knowledge silos weaken; hierarchies flatten.
This is where the industry veteran turns critical of media in general. "Campaign is also the problem, because everyone wants to get their interview or photo on it. That's why our industry is not as trusted as the investment banking or the consulting industry."
The real constraint, he says, is politics or the inability of people to work together. He feels there's an inverse relationship between willingness to cooperate and one's place in an organisation; the higher up one goes, the less cooperation they get.
At Monks, the operating matrix is country- or region-first, then client, then capability. But the downside of a country-first approach, he concedes, is far less than the downside of a discipline-first approach, because that leads to people fighting with one another.
From 'wow' to 'show'
Sorrell frames his career in three acts: Saatchi & Saatchi as globalisation, WPP as its continuation, and S4 as fragmentation plus technological efficiency at scale. Geography and technology have always been twin forces; the balance has shifted toward the latter.
Consumers, he argues, are adopting AI faster than companies. Automotive and financial services feel external pressure, from Chinese EV makers like BYD to fintech disruptors. CFOs, not CMOs, may become the catalysts for change, especially as creative costs creep above an efficient 10% of media spend.
Fear remains. "It's like turkeys don't vote for Christmas." When he worried aloud to JP Morgan's futurist Deborah Kantt about AI's impact, she replied, "If you can give me another five or 10 years, I might be able to give you another 25 years." Longevity may increase; displacement anxieties will too.
Sorrell's shorthand for AI's evolution is telling. "In 2024 with AI, we said, 'wow'; in 2025, it was 'how'; in 2026, it is 'now'." Could 'show' be the big word in 2027? Sorrell feels it well might, because it means that Monks' clients will operate in a much more effective and efficient way. AI, he says, will raise the baseline of creative output. "We'll get rid of the crap, and there'll be a much higher level of super creative stuff produced by human beings."
The unresolved question is not whether machines will improve the average. It is what happens, as he puts it, "when machines are better than people." For agencies navigating frenemies, fractured correlations and flattened hierarchies, that is less a philosophical puzzle than a commercial one.
In early February 2026, Wall Street was reminded how fragile conviction can be in the age of artificial intelligence. When Anthropic released plugins for its Claude Cowork AI agent, nearly USD 300-billion in market value evaporated in about a single trading day. Investors were not reacting to a product update; they were repricing entire business models.
For Sir Martin Sorrell, the sell-off was a referendum on structural uncertainty. "What the market is trying to say is there's this significant thunderstorm of uncertainty. People are worried about AI's impact on the traditional holding company business model," he told Campaign.
He points to the divergence among the large groups. "WPP is talking about minus 8% in Q4 of last year, while Publicis was plus five, which is 13-point spread between two of the biggest agencies."
The contrast, he argues, reflects strategy and structure. "Publicis has a strong strategy and leadership with Maurice Levy handing over to Arthur Sadoun (CEO of Publicis Groupe). We have to see whether it can deliver in practice, particularly around data and digital with Epsilon and Sapient. They have the best structure, which Monks mirrors, which is country-first, client-second, and discipline or capability, third. Omnicom has gone the other way with discipline-first, which is a recipe for disaster. WPP looks as though they're going the same way, which creates too much internal friction."
In July 2018, Sorrell's S4 Capital acquired Dutch digital agency MediaMonks for approximately USD 350-million to establish a new-age digital marketing services firm. In 2024, it was rebranded to simply Monks to emphasise a unified, AI-driven, and talent-focused approach.
K-shaped industry, a broken correlation
Sorrell describes the communications industry as K-shaped. "It's USD 1.2-trillion (global media market) of which USD 900-billion is digital, growing at 15-20%. The USD 300-billion, which Warner Bros Discovery and Paramount are scrapping over, is going backwards. So free-to-air (FTA) TV is under huge pressure. Even in India, where live sport buttresses it, it's still difficult," he added.
That split has fractured a once reliable metric. He maintains there's almost a one-to-one correlation between corporate profitability and ad agency revenues, which he believes is broken down. "Firstly, because traditional business is going backwards, which is where the heart of Omnicom, Publicis, WPP and Dentsu is based. And then you've got the digital part, which is growing like Topsy," he said.
At the same time, the hyperscalers are redirecting capital. Alphabet, Meta, Amazon and others, are reducing their marketing experience, according to Sorrell, because they're spending altogether over USD 600-billion this year on data and AI infrastructure, data farms, and more. He opined that they're switching their spending from Opex marketing to capex and that the correlation between company profitability and agency revenues or advertising spending has broken down.
Trade needs to be sharper and smarter
Sorrell says, "I think Campaign magazine and other media should report the digital industry and the traditional industry separately." Aggregating them obscures where growth and decay actually sit.
The digital half of that USD 1.2-trillion pie is highly concentrated. Sorrell states that marketers basically have six choices — Alphabet, Meta, Amazon, Alibaba, Tencent and Bytedance. He lists their contribution: of this, Google this year probably has USD 300-billion, Meta has USD 200-billion, Amazon's about USD 70-billion, and TikTok around USD 50/60-billion. With hyperscalers spending USD 600-billion on AI capex, he believes its market share will grow, which also makes the entire landscape more concentrated.
However, this concentration creates dependency. "But agencies have to get closer to the platforms like frenemies." He recalls using the term with former Google executive and current Palo Alto Networks CEO Nikesh Arora at Cannes.
The threat is not theoretical. He doesn't think the day is far when clients will be able to automatically create, produce, plan, buy, and distribute advertising campaigns with an end-to-end tech model. Especially for a server message block (SMB) customer, who can merge Veo 3, Nano Banana and Performance Max — or Advantage Plus in the case of Meta. Platforms publicly downplay enterprise ambitions, he argues, because "they don't want to alienate agencies, because it is a direct attack."
Behind market jitters is another anxiety. This could be due to the fear that the media model will come under attack because the technology exposes all the frictions and inefficiencies in the media chain.
"We all know there are sticky fingers in media discounts, rebates, and more, which maybe don't accrue to clients. But in a fully pre-automated, fully mechanised system, these frictions will disappear," Sorrell predicted.
In that world, agencies must redefine their role. Algorithms will decide the other 80%, but somebody has to independently decide whether the algorithm works, which is where agencies step in. Validation, arbitration, and strategic counsel become core. According to Sorrell, that means that the agency model has to shift dramatically from what it is.
He notes that AI has yet to fundamentally rewire some incumbents. "Omnicom and IPG have done nothing with AI in the last year; they've dabbled with it, but AI hasn't fundamentally altered how they operate." Yet headcount has already fallen from 127,500 to 104,000. "The real headcount reduction is yet to come."
In today's times, strategy decks, Sorrell implies, will not suffice.