Loading...
Two agencies can do almost everything right with AI and get opposite results. Omnicom grew 6.7% while WPP fell 6.6% — the real split is which layer they deployed AI at.
The holding company earnings season just revealed something the industry doesn't want to admit: two agencies can do almost everything right with AI and get completely different results. Omnicom reported Q1 2026 revenue of $5.6 billion, up 6.7% year-over-year, with a plan to offload $3.2 billion in businesses and merge or sunset more than 20 agency brands. WPP reported £3,030 million, down 6.6% year-over-year, with North America down 7.8% — despite raising $600 million in debt to fund AI partnerships with Google and Adobe. The divergence isn't random. It's a structural split between two different theories of what AI is actually worth.
One theory says AI is an execution multiplier. Deploy it across media buying, content production, influencer workflows, and data processing, and you compress costs while scaling output. The other theory says AI is a strategic accelerant. Deploy it to synthesize market intelligence, compress decision cycles, and give strategists more time for judgment-heavy work. Both theories are partly right. Only one is generating growth.
The most telling detail from the Omnicom results isn't the growth — it's what they're growing with. Their new agentic AI tool, Creo, uses Google Gemini to automatically fix brand safety issues in creator content. Their focus is on integrated media, commerce, data, CRM, consulting, and content automation. Everything on the execution layer.
When you look at what holding companies are actually building, the pattern is clear. WPP's $600 million bond issuance funds WPP Open, a media and production platform. Omnicom's restructuring targets integrated media execution. Both are betting that AI-powered execution at scale is the differentiator. And here's the uncomfortable truth: they're probably right about the market in the short term. According to Gartner, 40% of enterprise apps will feature task-specific AI agents by 2026, up from less than 5% in 2025. The tools for execution-layer AI are becoming commodity infrastructure.
That's exactly the problem for agencies that bet everything on execution. When your differentiation is faster content production, cheaper media buying, and automated workflow tools, you're competing in a space where barriers to entry collapse every 18 months. The same AI capabilities Omnicom is deploying internally will be available to any competitor, any holding company, any well-funded startup. Execution efficiency is a treadmill — you run faster just to stay in place.
Meanwhile, the question that actually determines whether companies win or lose — which markets should we compete in, how should we position, where should we allocate resources — that question is getting less attention, not more. Gartner predicts over 40% of agentic AI projects will be canceled by end of 2027, with organizations "blind to the real cost and complexity of deploying AI agents at scale." The failure rate isn't primarily technical. It's strategic. Teams are building AI infrastructure for goals that were never clearly defined.
This is where the holding company bifurcation reveals its deeper meaning. WPP's decline isn't a failure of technology adoption. They have partnerships with Google and Adobe, significant AI investment, and a platform strategy. The decline is a failure of strategic clarity — the work of helping clients understand which moves to make and why. That judgment-heavy, synthesis-heavy, decision-ready work is exactly what AI can accelerate when deployed as a strategic partner rather than an execution tool.
The irony is that the agencies growing by focusing on execution are building exactly the capabilities that will be most commoditized. The agencies declining despite significant AI investment are declining not because they adopted AI too slowly, but because they adopted AI for the wrong layer. Tool sprawl and execution-layer accumulation are different expressions of the same problem: strategic clarity getting lost beneath operational activity.
The divergence between Omnicom and WPP isn't permanent, and it's not a verdict on AI strategy. It's a snapshot of a market that hasn't finished sorting. Right now, execution efficiency commands premium valuations because it's measurable and attributable. Media optimization produces measurable ROAS. Content production produces measurable output. The strategic clarity question — are you making the right bets? — produces outcomes that take quarters or years to evaluate.
That's why the holding companies are bifurcating. Execution is easier to sell, easier to measure, and easier to scale through AI. Strategy is harder to attribute, harder to commoditize, and harder to deliver consistently. The agencies growing today are the ones that found a repeatable, scalable execution product. The agencies declining are the ones that haven't — regardless of their strategic sophistication.
But the sorting isn't finished. Deloitte's 2025 Tech Value Survey found that 74% of surveyed organizations invested in AI in the prior 12 months, yet the same research shows a persistent gap between AI investment and ROI. Organizations are spending heavily on AI infrastructure without a clear line to strategic outcomes. When that gap becomes impossible to ignore — when boards start asking why AI investment isn't producing competitive advantage — the market will swing back toward strategic clarity as the differentiator.
If you're a CMO right now, the holding company divergence is a warning sign disguised as good news. Omnicom's growth looks like validation that AI investment works. WPP's decline looks like a cautionary tale about falling behind. The reality is more nuanced: both outcomes reflect investment in execution-layer AI, just at different stages of market saturation. Neither outcome tells you whether you're asking the right strategic questions.
The question to ask any agency partner — holding company or independent — isn't whether they use AI. It's which layer they're deploying it at. Execution-layer AI produces faster, cheaper, more scalable output. Strategic-layer AI produces better decisions, clearer priorities, and more confident resource allocation. Most agencies can demonstrate the first. Few can consistently deliver the second.
When you're evaluating partners, ask specifically: where does AI appear in your work? If the answer centers on content production, media optimization, workflow automation, or data processing, you're looking at an execution partner. If the answer centers on synthesis, decision-making, market positioning, and strategic judgment — if the AI is doing the analytical work that lets your team focus on the decisions that matter — you're looking at a strategic partner. Both have value. They require different expectations, different timelines, and different definitions of success.
Here's what the holding company divergence actually reveals: the sequence of investment matters more than the investment itself. Strategic clarity before execution efficiency produces compounding returns. You make better bets, allocate resources more effectively, and build execution muscle that's pointed in the right direction. Execution efficiency before strategic clarity produces expensive momentum in the wrong direction — faster, cheaper output that compounds strategic misalignment.
Most organizations are discovering this the hard way. According to research compiled from multiple sources, the productivity benefits of AI are real but unevenly distributed, with organizations struggling to measure and optimize adoption patterns across teams and roles. The organizations capturing the most value aren't the ones with the most AI tools or the highest AI spending. They're the ones that aligned AI investment with strategic decision-making before scaling execution.
This is the underlying logic of the market sorting happening right now. The agencies growing through execution efficiency will continue to grow until the execution layer saturates. The agencies that invested in strategic clarity alongside execution will be better positioned when the market pivots back toward judgment. The question isn't whether AI works. It's whether you're using AI to execute faster or to decide better.
Book a 30-minute demo. Bring a live question and watch the answer get built.