As AI revolutionises marketing, the traditional agency model is undeniably under threat. In this article, Tim Hussain, co-founder of Gen AI consultancy Signal42, and former advertising leader at OLIVER, Ebiquity and Sky, explores how Gen AI is exposing structural flaws in holding companies, and what agencies must do to survive.
For the better part of almost a century, big advertising agencies had an outsized influence on marketing. They were the trusted custodians of brand image and the maestros of mass media. Their holding company parents – WPP, Omnicom, Publicis, IPG and the like – grew into billion-dollar empires, buoyed by steady fees and a lack of serious competition. That era is swiftly drawing to a close. Artificial Intelligence is set to erode the traditional agency model, and the cracks in the foundation are already turning into chasms.
Consider the writing on the wall. Clients have been losing faith in the old model for years. Many built in-house teams to regain control and cut costs – some 82 percent of major advertisers now have in-house agencies, up from just 42 percent in 2008 . Consultancies muscled in, offering something agencies didn’t: direct linkage of marketing to business outcomes. While agency holding companies were busy rearranging their org charts or merging subsidiaries to “synergise,” Accenture built a marketing services unit that hit $18 billion in revenue by 2023. Advertising’s incumbents have been playing defence.
Now along comes AI – the great disruptor. We’re not talking about incremental automation; we’re talking about a step-change in how marketing work gets done. Generative AI can produce in seconds what an agency might take weeks to deliver. Need a dozen new ad visuals? An AI like DALL-E can generate them instantly. Copy for a 30-page brochure? GPT-4 can draft it overnight. Media plan optimised across five platforms? Let an AI loose on the data and it will allocate budget more efficiently than a roomful of media planners. OpenAI’s CEO Sam Altman projects that “95 percent of what marketers use agencies, strategists, and creative professionals for today” could be handled by AI at near-zero cost . Even if that figure ends up a bit high, the direction is clear.
The Agency Value Proposition, Undone
Traditional agencies have long sold themselves on a combination of creative excellence, strategic counsel, and executional scale. AI chips away at each of these pillars:
Executional Scale: Holding companies grew big by offering armies of talent – a global footprint, hundreds of specialists, the ability to “do it all” for a client. But scale can quickly turn into a liability. Large agencies have massive overheads. AI flips the script: a small firm (or an in-house team) with powerful AI tools can execute campaigns that might have required dozens of people in the past. It’s very feasible to imagine a lean “virtual agency” of, say, five people who manage AI systems to output creative, place media, and optimise – servicing a multi-million-dollar account that used to feed an agency office of 50. This isn’t sci-fi; it’s the logical extrapolation of tech trends already underway. No wonder industry vets are speculating about “a new breed of agency that is people-light and technology-adept”.
Creative Excellence: We’ve comforted ourselves that creativity is a human domain – a machine couldn’t possibly match Don Draper’s ingenuity, right? The reality is more nuanced. AI may not feel emotion or truly understand culture, but it can rapidly iterate creative options at a volume and speed humans can’t. Already, AI-generated art and copy are passing Turing-esque tests. The “creative” output across much of advertising is mediocre anyway (let’s be honest), and as David Droga bluntly puts it, “the majority of our industry is bullshit… That’s why people invented technology to try and either block it, or do it better.” In other words, a lot of advertising isn’t some sanctified art – it’s filler. And if it’s filler, a well-trained AI can do it. The bar for human creative input will rise to the truly original, insightful big ideas. How many agency networks are consistently delivering those versus churning cookie-cutter social posts and banal banner ads? AI is going to expose just how much of agency work was routine production dressed up as creativity.
Strategic Counsel: Top agencies love to claim they are “business partners” and “brand stewards” with deep expertise. But when was the last time your ad agency told you something about your consumers that you didn’t already know? Agencies often lack the hard data and analytics chops that consultancies or specialised data firms have. Now AI comes along, with the ability to synthesise vast consumer data, market research, and competitive intel in moments. It can write a situation analysis or even suggest a marketing strategy blueprint if fed the right inputs. The strategist armed with AI will outperform the strategist without one – and a client armed with AI might not need an agency for insight at all. As Michael Farmer, author of Madison Avenue Manslaughter, observed, agencies rarely ask fundamental business questions; they focus on making the ads.
Exposing the Structural Flaws
AI isn’t arriving in a vacuum; it’s pouring fuel on a fire of issues that have long been smouldering in ad-land . Chief among these is the broken financial model of agencies. Most holding-company agencies still bill clients largely by hours (or FTEs) and mark-ups. This model has been under siege for decades – procurement ensured that agency fees got compressed and any fat was trimmed. Agencies coped by cutting costs, which usually meant juniorising teams and increasing volume of work to make up for lower fees per project.
The result: overworked staff, mediocre outputs, and a transactional client relationship. Enter AI, which promises to automate much of the “toil,” as analyst Josh Chasin notes. Great, right? It frees humans to focus on higher-value aspects. But here’s the rub: if the old model was charging for head-hours, and now an AI does the work in a fraction of the time, who pays for that? Unless agencies overhaul how they charge – moving to value-based pricing or outcomes – they’ll simply end up doing the same work for less money. In effect, AI could accelerate the commoditisation of agency services that is already far advanced.
Another flaw is the holding company structure itself. These conglomerates have collected dozens of agency brands under one roof, but too often that has resulted in internal turf wars, duplication, and bureaucracy. Clients frequently complain that different parts of the same holding group don’t talk to each other or pitch in cohesive ways. WPP and others have tried to solve this by merging agencies or creating “Power of One” teams for clients, but these are band-aids.
The truth is, holding companies were a financial innovation (roll up agencies, gain bargaining power and scale) that hasn’t delivered commensurate innovation in output. As Farmer quips, aside from consolidating financials, do holding companies really add value? It’s an open question. And if the stock market starts doubting the answer, watch out. Because investors have no sentimental attachment – if growth stalls and margins erode, holding companies could face calls to break up, just like the bloated conglomerates of yesteryear . AI’s rise could be the catalyst that makes everyone realize “what the hell is a holding company anyway?” if each component agency could theoretically partner with tech firms or private equity and do better unshackled.
Adapt or Fade Away
I’m not writing an obituary for the big agencies just yet. There is a path forward, but it requires swallowing some tough medicine and embracing change with a zeal we’ve yet to see from many of the incumbents. Here’s what needs to happen:
1. Radically re-imagine the offering: Agencies have to stop selling advertising outputs and start selling business outcomes. If AI can generate a campaign, the agency’s role should be ensuring that campaign truly moves the needle for the client’s business. This might mean focusing on higher-level advisory, brand strategy infused with data, or creative platform ideas that an AI can execute across channels. It definitely means tying fees to results. As one Accenture Song exec said, trying to preserve the old fee model is “standing on a shrinking iceberg” . The iceberg is melting fast.
2. Get serious about tech (finally): For years agencies outsourced a lot of tech and data work or bolted it on via acquisitions. Now, every agency leader should be immersing themselves in AI capabilities and data infrastructure. It’s telling that Publicis is spending only ~$1000 per employee on its big AI program, and WPP about $3,000 per employee – those numbers pale in comparison to true tech companies’ R&D spend (OpenAI and friends are spending billions on AI research ). If agencies don’t invest deeply in their own AI tools and training, they will always be a step behind, reliant on outside vendors or clients’ tech – and thus easily replaceable.
3. Slim down and sharpen up: This is heresy to holding company chiefs, but it’s time to streamline. Maintaining a constellation of half-performing agency brands is a luxury of a bygone era. The future belongs to focused expertise. Some holding companies might need to merge creative, media, and digital into one strong offering (honestly, clients want integrated solutions, not a different logo on every PowerPoint). Others might spin off or close units that don’t have a clear value-add in an AI-enhanced future. It’s better to cannibalise parts of your own business on your terms than to watch a tech startup do it for you.
4. Champion creativity + AI, not creativity vs. AI: The industry often frames it wrong: human creativity versus artificial intelligence. The winners will be those who figure out human creativity augmented by AI. That means celebrating the creative teams that leverage AI to be 10x more productive or to inform their ideas with richer insights. It means case studies where an agency used AI to do something truly novel and effective – not just to cut a few hours from the process. Imagine a brand campaign that reacts in real-time to cultural moments (via AI trend analysis) and deploys new creative variations on the fly. That’s the kind of agile, AI-fueled creativity agencies should offer – the kind no in-house team has mastered yet. Use the technology to push boundaries, not just trim costs.
At the end of the day, will the holding groups rise to this challenge? Or will they slide into irrelevance as AI democratises capabilities and clients seize the reins? I’ll be candid: the track record of these incumbents is not inspiring. Change in this industry often happens slowly, and usually only after major client losses or financial pain. Well, that pain is arriving. The next 2-3 years will be telling. Some holding company CEO will inevitably get on an earnings call soon and blame “AI-driven client insourcing” for a revenue miss – mark my words.
The hopeful view is that this very threat catalyses overdue change. Fear of death can concentrate the mind. Perhaps we’ll see an unexpected renaissance where agencies prove they can deliver more value than ever by combining human ingenuity with machine efficiency. I actually believe a few will manage it – those with bold leadership unafraid to break the mold. But many others will likely flounder, doubling down on yesterday’s formulas in the vain hope that a “human touch” alone will save them. It won’t.
The old advertising holding company model is living on borrowed time. AI didn’t start the fire that’s burning it down – that started when agencies forgot that they’re supposed to solve business problems, not just make ads. But AI is certainly fanning the flames. Adapt, reinvent, or get out of the way. The market has no sympathy for those who cling to the past while the future barrels down upon them.
