Tom Amies-Cull
23 Apr 2026
This moment will separate the builders from the cutters. The window is narrowing.
The Money Is Coming Out. The Value Is Going With It.
This is one of the most genuinely interesting moments the agency industry has faced in a generation. The tools are real. The disruption is real. Client expectations are shifting, commercial models are under pressure, and the question of how knowledge-based organisations create and deliver value is more open than it has been in decades. That is a difficult set of conditions to navigate. It is also a genuine opportunity to build something better. The businesses that recognise this and act on it now will define what the industry looks like in five years. The ones that do not will have optimised their way to the wrong destination.
What is actually happening in most cases is simpler and less flattering. AI is being used as the justification for decisions that were already going to be made. Headcount is coming out. Costs are being restructured. And the whole thing is being packaged as forward-thinking transformation. When Omnicom doubled its labour cost-savings target from the IPG acquisition, from $750m to $1.5bn, its share price jumped more than 13 percent the following day. WPP has followed: 7,000 roles cut, £500m in savings targeted by 2028. Analyst Ian Whittaker put it plainly in Campaign: AI is not driving this. It is the cover story. The driver is shareholder pressure on the cost base.
That might be a defensible short-term response to shareholder pressure. It is not a strategy. And the cost of confusing the two is going to be significant.
The productivity gap nobody wants to talk about
Alongside the restructuring story sits a data point that should be making every senior leader uncomfortable.
As reported in Fortune this week, a study published earlier this year by the National Bureau of Economic Research, covering around 6,000 executives across the US, UK, Germany and Australia, found that nearly 90 percent of firms said AI had no meaningful impact on productivity over the past three years. Average usage among senior leaders: 1.5 hours a week.
Economists are calling this Solow's productivity paradox repeating itself. Robert Solow identified the original version in 1987, when computers were everywhere but refused to show up in the productivity numbers. Apollo's chief economist made the same observation about AI recently: the value depends not on the tool itself, but on how it is actually deployed across an organisation.
The tool is not the transformation.
Research from BCG in a recent HBR article on AI and 'Brain Fry' adds a practical edge to this: workers using three or fewer AI tools reported productivity gains, but those using four or more reported cognitive overload and more errors. More tools, layered onto unchanged structures, creates friction rather than performance. And yet the dominant industry response is to cut headcount, purchase AI licences, and announce a transformation programme.
Who is winning and why
Publicis seems to stand apart. Twenty consecutive quarters of growth, with organic revenue up 5.9%in Q2 2026. CEO Arthur Sadoun has been direct that the business is not chasing the Wall Street cost-cutting playbook. He can say that partly because of how the business is structured. As Whittaker points out, The Badinter family's voting rights give Publicis genuine insulation from short-term shareholder pressure, and the freedom to choose a longer horizon. Most listed rivals do not have that option.
More instructive comparators sit elsewhere. Independent agencies, less visible and not beholden to holding company reporting, are winning pitches, holding margin, and building commercial models that do not depend on billing hours. When the leadership team is also the delivery team, strategy and execution stay close together.
Private equity-backed businesses tell a related story. PE is often painted as short-termist, but in practice these businesses tend to have clearer value creation mandates, tighter operating discipline, and critically they are not resetting priorities every 90 days to manage analyst expectations. When they deploy AI, it tends to serve a specific commercial goal rather than an efficiency narrative.
For listed businesses without these structural freedoms, the lever available is narrative discipline: building a credible case for where growth is actually coming from, not just how much cost is coming out, and holding that position through reporting cycles. Most are not doing this. Some could.
The real problem: stuck in old ways of thinking
Four things explain why the current industry response falls so short, and one condition connects all of them.
First: treating AI as a cost tool rather than a growth tool. Cutting headcount using AI as the rationale is the obvious short-term move. The harder question, where does AI help us build something clients will find critical to growing their own business, might get asked but rarely gets planned and invested seriously because it requires a longer view than quarterly results tend to allow.
Second: avoiding the commercial model conversation. The billable-hour model is not just under pressure. It is broken. And cutting heads does not fix it. It changes the shape of the P&L without solving the underlying problem. Replacing human capacity with AI capacity is not a straightforward saving either. Technology infrastructure, licensing, integration and governance all carry real costs, and unlike people, not all of those contracts do not flex easily when revenue dips. The businesses getting ahead of this are moving toward value-based pricing, outcome-linked retainers, and revenue lines built on IP and proprietary capability rather than hours and headcount.
Third: deploying AI without redesigning how work actually happens. AI layered onto unchanged structures, workflows and incentive systems produces noise rather than performance. You’re digitising dysfunction. The real question is not which tools to buy. It is how the organisation needs to work end-to-end to create value for clients, and where AI genuinely enables that rather than just automating a broken process.
Fourth, and most underestimated: culture. Not as a soft add-on, but as the thing that determines whether any of the above lands in practice. Confidence in AI among workers fell 18 percent in a single year even as usage increased. In a complex, multi-generational workforce navigating real uncertainty, that trust gap does not close through town halls or training decks. It closes through honest leadership, genuine investment in people's development, and a culture visibly connected to commercial outcomes rather than layered over structural anxiety. Culture that is real creates the trust, adaptability and shared direction that make transformation stick. Culture that is theatre accelerates the problem.
The thread connecting all four: these are not separate failures. They are symptoms of leadership teams optimising an old system rather than building a new one.
It is worth being clear about where this is most acute. The failures described here sit predominantly in the front and middle office: the capabilities, processes and ways of working that directly deliver client work. That is not to diminish the support functions that make it all possible. Finance, HR, technology, operations: the same absence of vision and real ambition to reimagine how work gets done runs through those teams too, and deserves its own conversation. But the commercial consequences are most immediately visible at the client-facing layer, and that is where the urgency is greatest.
What actually building something looks like
Three things consistently separate organisations that are genuinely transforming from those going through the motions.
Start with what clients will actually need, not what you currently sell. Before any tool is deployed or team restructured, the real question is: what problems will our clients face in two to three years, and are we building toward solving those now? The answer requires looking outward, at how clients' own businesses are being disrupted, how their customers are changing, and what that means for the capability and insight they will need from partners. The agencies that will grow are those ahead of their clients' next challenge. Those that will struggle are still solving the last one.
Get honest about the commercial model. Does the way you charge for work actually reflect the value you create? For most agency businesses, the honest answer is no. The shift required is from pricing inputs, hours and headcount, toward pricing outcomes. That also means modelling the full economics of AI deployment: what does the technology infrastructure actually cost to replace the headcount it displaces, and how does that cost base perform across different revenue scenarios? The businesses navigating this well are having that conversation openly rather than announcing net savings that obscure a more complicated picture.
Make culture the foundation, not the final slide. Strategy and operating model redesign can be intellectually correct and still fail if the culture conditions are not there to support them. That means being honest with people about what change means for them, investing in their ability to grow into it, and connecting day-to-day work visibly to commercial outcomes. In a workforce spanning multiple generations with varying relationships with technology, this is the hardest leadership challenge right now. It is also the one that most determines whether transformation actually delivers.
The question worth asking
Ian Whittaker closes his Campaign piece with a sharp question: not whether shareholders will tear the industry apart again, but who gets torn apart next.
It is the right question for investors. For the leaders of these businesses, a more useful one is this: are you going to be builders, or cutters?
For now, yes. McKinsey’s Corporate Horizon Index, tracking over 600 large companies across 15 years, found that businesses with a genuine long-term mindset consistently outperformed their peers across almost every financial measure that matters - and while older research, it has never more relevant! Research published through Harvard Law School’s corporate governance forum was starker still: short-termist pressure leads to a collective race to the bottom, where firms fail to invest in exactly the projects that would have driven the most long-term value. The pattern is well documented. The agency industry is living it right now. But the research also shows that the companies which recognised the trap and changed course did pull ahead. That is the real story, and it is still available to those willing to act on it.
I’m sure this current restructuring wave will be studied by future MBA students not as a transformation story, but as a case study in what happens when short-term financial pressure fills the space where long-term strategy should be. The more interesting case studies will be the businesses that looked at exactly the same disruption and made a different choice: to redesign their commercial model, rebuild how work gets done, and invest in the culture that made it stick. Those businesses are making that choice right now. The window is open. The question is who goes through it.
The execution gap has always been a leadership gap in the end. The tools are not the problem. The ambition is. And ambition, at least, is a choice
I provide independent consulting, advisory and fractional leadership through IntelioWorks - supporting & enabling senior leaders on strategy, transformation, operating model design, AI readiness and culure. I also work with partners across the data, tech and transformation ecosystem to help create solutions that are relevant, agile and additive for businesses whatever their scale and maturity.
Article first published on LinkedIn April 23 2026