Agencies in the AI Shake-Up: The Market Is Booming – and the Middle Is Getting Squeezed

The advertising and agency industry entered 2026 divided. One camp expected tangible productivity gains from AI, the other pointed to open questions around data, quality and cost. There was no talk of euphoria.¹ Now the half-year numbers are in – and they prove both camps right, just not at the same time: the market is growing strongly, but the gains are spread extremely unevenly. To understand the AI shake-up of agencies, you have to read both movements together. And the lesson reaches far beyond the industry.

The Expectation: A Divided Industry

Both camps could point to numbers. The sceptics had the McKinsey study "State of Marketing 2026" on their side: among roughly 500 marketing executives surveyed across the five largest European economies, artificial intelligence ranked only 17th in priority, and 94 percent said they had not yet made any meaningful progress in AI implementation. The top 5 were dominated by classic core disciplines such as brand building and authenticity.² The optimists could cite the same study: companies that had already built out their AI capabilities put their efficiency gains at around 22 percent.² The expectation was not naive but split in two – AI could move a lot, but only for those who went about it in a structured way.

The Reality: The Market Is Growing – with AI Named as the Driver

First the good news, and it is unambiguous: the market data shows no sign of an agency die-off. The global advertising market is growing by 8.9 percent to around 1.3 trillion US dollars as of mid-2026, well above the 7.1 percent forecast at the turn of the year. The named drivers are AI-native advertisers as well as traditional companies using AI to increase efficiency.³ The forecast comes, of all places, from the house of WPP – from the media arm of a group that is itself in the middle of a restructuring. The more cautious Dentsu forecast also sees a plus of 5.0 percent and expects around 75 percent of global ad spend to be algorithmically driven by 2028.⁴

So AI is not destroying the industry – on the contrary, the forecasts list it as the central growth driver. But this growth is not evenly distributed. And that is precisely where it is decided who benefits from the shake-up.

The AI Shake-Up Splits the Market

The 2026 reporting to date from the two large European holding companies shows two opposing trajectories (Publicis has published half-year results, WPP so far the first quarter):

The decisive point sits behind the numbers. Publicis has spent years rebuilding its business around data and AI – from audience targeting through campaign steering to performance measurement. AI is not sitting in the shop window as a product; it is embedded in the service itself and makes it better: more precise audiences, more measurable results, more impact per advertising euro. That is what clients pay for – with the result of 4.8 percent organic growth in the second quarter and a raised full-year guidance.⁵ How deep the rebuild goes is shown by one figure from the half-year report: 87 percent of net revenue falls, by the group's own categorisation, under "AI-powered marketing services", from data through media to production.⁵

WPP stands at the other end: in the first quarter of 2026, net revenue fell 6.7 percent like-for-like, and the group expects a mid-to-high single-digit decline for the first half.⁶ WPP names the causes itself: in its own strategy update, the group speaks of "excessive organisational complexity, the lack of an integrated operating model and inconsistent execution". The consequence is a far-reaching restructuring: under the "Elevate28" programme, the holding company is being merged into a single, AI-integrated company; by 2028, WPP aims to cut around 500 million pounds in annual costs (gross) and redeploy part of that into AI, data and growth.⁶ The rest of the market is realigning too: at the end of 2025, Omnicom and IPG completed their merger into the world's largest agency group.⁷

The weakness of any single group always has several causes – at WPP, client losses, a weak media business and leadership changes came together. The pattern across the industry is nevertheless clear: out in front are the houses that have combined data, AI and creativity into one integrated offering. It is not access to AI that decides – everyone has that – but the structure it is embedded in.

Three Positions – and Why the Middle Is Getting Squeezed

The same dividing line runs through all size classes. It does not run between big and small, but between clear positioning and interchangeable execution. Three positions can be distinguished. At the top end are the large integrated platforms; they win through scale, data and AI – through reach, measurability and a one-stop offering, for which Publicis delivered the clearest example this half-year. At the bottom end are the small specialists that win through depth and speed: lean, highly specialised AI-first teams that deliver faster and more precisely in a clear niche than traditional shops. In between sits the middle – the classic full-service agency with typically 30 to 150 employees that does "a bit of everything" and competes primarily on output. That is exactly who the HORIZONT column titled "The Middle Is Dying" means.⁸

The middle is coming under pressure from two sides at once. From below, generative AI makes cheap precisely what its core business consists of – producing copy, images and campaigns. From the client side, the simple, recurring tasks that remain are increasingly moving into in-house marketing departments.⁹ So the mid-sized generalist is too big and too expensive to be as agile as a boutique – and at the same time not differentiated enough to keep up with the data and AI strength of the large platforms. The problem lies in the business model: selling execution is losing value.

The way out leads to one of the two ends of the scale. For smaller providers, that is actually good news: those who occupy a clear niche, consistently integrate AI into their own workflows and act as advisers can hold their own against much larger competitors. Our guide AI Consulting for Agencies describes what a structured start looks like for agencies. To begin with, an honest self-assessment is enough: are we mainly selling output – or an advantage a client cannot simply reproduce with AI themselves?

Why This Also Affects Companies Outside the Industry

The obvious objection: Publicis and WPP are billion-euro groups; their problems are far removed from the day-to-day of a company with 80 or 300 employees. But the mechanism does not depend on company size. Agencies sell exactly the services that generative AI makes cheap first: copy, images, campaign steering. That is why they feel the price pressure earlier than other knowledge-based service providers – earlier than law firms, consultancies or engineering offices. What is written in agency balance sheets today shows those industries, with some lead time, where things are heading: paid execution is losing value, integrated advice is gaining.

A second finding from the half-year reporting fits this. The trade publication Adweek sums up the Publicis numbers like this: clients are adopting AI but postponing the fundamental transformation of their business.⁵ This gap between tool usage and real restructuring is exactly the gap McKinsey measured at the start of the year – 94 percent without meaningful progress, while those proceeding in a structured way gain 22 percent in efficiency.² The thread comes full circle: using AI is not enough. What matters is reorganising the business around it.

What the AI Shake-Up Means for Your Company

A simple but uncomfortable consequence follows: efficiency gains from AI are becoming the standard and will soon reach every competitor. An edge therefore comes not from the tool but from what is hard to copy – clear strategy, clean data, an unmistakable positioning and a brand with judgement. For agencies, that concretely means: those who mainly sell execution come under price pressure; those who design processes and orchestrate systems become strategic partners. For mid-sized companies, the same logic applies – except that, unlike the agencies, they still have some lead time to act.

Recommended Actions: Three Steps

1. Audit your value creation. List your ten most important services and mark every one that a client could handle themselves with an AI tool in the foreseeable future. Everything marked is your margin risk. The rest shows where your future lies. A structured AI assessment provides the baseline for this.
2. Bring AI into one core process. Publicis' edge comes from integrating data, AI and creativity into the service itself – not from buying yet another tool. Choose one core process, define data, owners and a quality standard, and see it through before starting the next. Introducing AI in phases shows how this works step by step.
3. Clarify your positioning. Test question: can your three most important clients say in one sentence what makes you irreplaceable? If not, that answer comes before any further AI project. Those who only restructure once the numbers collapse pay the higher price. Those who act beforehand secure their margin.

The AI shake-up is not decided by the technology but by the structure with which you steer it. That is exactly where we support you.

Frequently Asked Questions About the AI Shake-Up of Agencies

Will AI replace advertising agencies? The market data suggests otherwise: the global advertising market grows by 8.9 percent in 2026 – the forecast names AI as the most important driver.³ What comes under pressure above all is pure execution work; strategy, integration and brand management gain in value.

Which agencies benefit from the AI shake-up? The two ends of the scale: large providers with an integrated data and AI business, and small, specialised AI-first providers. Mid-sized generalists competing mainly on execution come under pressure from two sides.⁸

What does the shake-up mean for mid-sized companies? The agency industry shows early what is coming for knowledge-based service providers as a whole: efficiency gains from AI become the standard, competitive advantages arise from strategy, data and positioning.

Is introducing AI tools enough? The numbers suggest otherwise: 94 percent of European marketing executives report no meaningful progress despite available tools.² Impact only arises when AI is integrated into core processes with defined data and responsibilities.

Sources & References

1. HORIZONT, "Jahresausblick 2026: Die Werbebranche zwischen KI-Ernüchterung und Verantwortung", 14 Jan 2026: horizont.net
2. McKinsey & Company, "State of Marketing 2026" (approx. 500 marketing executives, five largest European economies), 21 Nov 2025: mckinsey.de
3. WPP Media, "This Year Next Year – Global Midyear Forecast 2026" (global ad revenue +8.9% to $1.3tn, excl. US political): wppmedia.com
4. dentsu, "Global Ad Spend Forecasts – May 2026" (+5.0% for 2026, 75% algorithm-driven by 2028), 27 May 2026: dentsu.com
5. Publicis Groupe, "First Half 2026 Results" (Q2 +4.8% organic, 21st consecutive growth quarter, 87% of net revenue "AI-powered marketing services"), 16 Jul 2026: publicisgroupe.com · Context: Adweek, "Publicis Boosts Full-Year Guidance as Clients Embrace AI, But Delay Transformation", 16 Jul 2026: adweek.com
6. WPP, "Strategy Update and 2025 Preliminary Results" (Elevate28, £500m gross annual savings by 2028), 26 Feb 2026: wpp.com · WPP, "First Quarter 2026 Trading Update" (net revenue −6.7% like-for-like, H1 outlook mid-to-high single-digit decline), April 2026: wpp.com
7. Omnicom, "Omnicom Completes Acquisition of Interpublic", 26 Nov 2025: investor.omc.com · Context: Campaign US, 26 Nov 2025: campaignlive.com
8. HORIZONT, "Die Mitte stirbt – was bleibt übrig? Was Agenturen jetzt nicht delegieren dürfen" (opinion column, Kim Notz, KNSK), 6 Jul 2026: horizont.net
9. HORIZONT, "Agenturen im Umbruch: KI oder K.O. – 2026, das Jahr der Wahrheit" (guest commentary, Sven Dörrenbächer, ex-CEO Antoni), 3 Mar 2026: horizont.net