Loading...
WPP's HEX launch is the highest-profile AI consultancy yet, but its accountability stops at capability and execution. The strategic decision layer stays unoccupied.
WPP just did something it has not done before. On June 5, the world's largest advertising holding company launched HEX, an AI-driven "production studio, R&D lab, and consultancy" that places roughly fifty creative technologists inside client organizations to build generative and agentic AI solutions, run AI workshops for Fortune 500 executives, and train internal teams. According to LBB Online, the unit is led by Elav Horwitz, WPP's first Chief Innovation Officer, and operates in partnership with Adobe, Google, and NVIDIA. The branding is the most direct positioning WPP has ever put on the word "consultancy" — and the substance is still execution-layer work. That gap is the most important thing a CMO will encounter in the next two weeks of Cannes noise.
HEX's own announcement places the unit inside WPP Production, frames its remit around generative and agentic AI workflows, gaming, immersive experiences, and robotics, and highlights early work like the SXSW 2025 "Dirt Is Good" activation for Unilever that used AI coaching and real-time motion capture. Decision Marketing's coverage notes the team draws from WPP's Creative Tech Apprenticeship and is measured on its ability to "embed forward-deployed experts with clients." That is a delivery-and-training model with a consultancy label attached. It is not a model that accepts accountability for the strategic decisions a buyer's business needs to make.
HEX is the highest-profile new entrant into a category that is suddenly very crowded, and very confused. The "AI consultancy" label now covers the launch of WPP's agentic-AI production unit, the founding of Europe-only "strategy meets agency" firms by former management consultants, Accenture Song's continued engineering-layer partnerships, the proliferation of "AI agencies" and "AI automation agencies" entering the market, and YC's W26 batch validating AI-native services as a category thesis. Every one of those firms is, in some sense, an "AI consultancy." Almost none of them are in the same business.
The reason the label has lost meaning is that consultancies have historically been defined by what they deliver. A strategy consultancy owns the recommendation. A management consultancy owns the transformation roadmap. A creative consultancy owns the campaign. The "AI consultancy" label is being applied to all of these and to several more — production studios, workflow integrators, and capability-training shops — without anyone having to commit to the deliverable. When every firm uses the same word, the buyer's question has to become structural: which layer of the work is this firm actually accountable for, and what is it not?
When a buyer hears "AI consultancy" in mid-2026, they are hearing one of three very different things, and the differences matter for what the engagement actually produces.
The first is the capability layer. These are the firms that train your team, build internal AI workflows, and embed practitioners who leave behind a working process. WPP's HEX sits in this layer. Accenture Song's partnerships with Google Cloud, ServiceNow, and HUMAIN sit in this layer. Their value is real and significant. Their accountability is for the working process they leave behind — not for whether your business made the right strategic decision.
The second is the execution layer. These are the AI production studios, creative shops, and campaign operators that deliver work. Their value is in the output. Their accountability is for the campaign, the asset, the launch. They are not in the business of telling you which campaign to run, which audience to prioritize, or which market to enter. The category is full of firms operating at this layer, and the layer is getting more crowded every quarter.
The third is the decision layer. This is where the strategic accountability lives — where a team takes ownership of the recommendation, defends it when the board pushes back, recalibrates when the market moves, and is named as accountable when the call turns out wrong. The decision layer is the layer most "AI consultancies" are not built to occupy, because it requires a fundamentally different model — continuous engagement, strategic ownership, and the willingness to put a name on the recommendation. It is also the layer that the buyer has the hardest time evaluating, which is why it is the layer most often left to trust and reputation rather than structure.
HEX is, by WPP's own positioning, a capability-and-execution layer consultancy. It is not, by structure or mandate, a decision layer consultancy. That is not a criticism — it is a different business. The problem is that the branding puts all three layers under the same label and lets the buyer sort out the difference on their own.
The reason this distinction matters is the data on what is actually happening inside buyer organizations. Deloitte's 2026 State of AI in the Enterprise report, based on a survey of 3,235 senior leaders across 24 countries, found that workforce access to sanctioned AI tools grew roughly 50% in a single year — and that only 25% of organizations have moved 40% or more of their AI experiments into production. The capability layer is succeeding: tools are being adopted, training is being delivered, and embedded teams are being built. The production layer is still where the gap sits. And the decision layer — the one that turns capability and production into P&L movement — is where most buyers have no clear owner at all.
Gartner's 2025 forecast projected that more than 40% of agentic AI projects will be canceled by the end of 2027, and the cancellation reason is not technology failure — it is the absence of clear strategic outcomes tied to the deployments. BCG's 2025 research found that only 5% of more than 1,250 surveyed firms are generating transformative AI value at scale. The same pattern appears in McKinsey's 2025 State of AI survey: 88% adoption, roughly 6% reporting meaningful EBIT impact. Capability is everywhere. Decisions are scarce.
The "AI consultancy" wave — WPP HEX, the strategy-and-agency hybrids, the YC-backed services companies — is a real and important response to a real capability gap. The training and the embedded teams and the production work are all valuable. But none of them close the decision gap on their own. The decision gap is structural, and structural gaps require a structural answer.
The reason most buyer organizations are stuck in the capability layer is not lack of investment. It is the synthesis tax. The average strategy team is running somewhere between six and twelve separate AI tools — each one producing fragments of insight that need to be integrated, validated, and translated into something actionable. The synthesis work falls on the team's highest-leverage people, who spend their time reconciling dashboard outputs instead of making decisions. The result is a stack of capability with no decision layer underneath it. The strategic work never lands.
This is exactly the pattern an "AI consultancy" at the capability or execution layer cannot fix. A training team can teach the team to use the tools. A production team can deliver work informed by the tools. Neither can absorb the synthesis cost of integrating six to twelve tools into a single defensible recommendation. The fragmentation is upstream of where they operate, and the buyer is left holding the bag. BCG's research calls this the "future-built" gap; we call it the synthesis tax. The label does not matter. The cost does.
If you are sorting through AI consultancy pitches this quarter, the structure of the question is straightforward. Three diagnostic prompts separate the three layers cleanly.
First, ask who owns the strategic recommendation this engagement produces. If the answer is "our team of consultants produces the analysis and your team decides," you are buying capability or execution, not decision accountability. If the answer is "we own the recommendation and we will defend it in the boardroom," you have found a partner in the decision layer. The first answer is the dominant one in 2026 pitches. The second is the one that actually moves the business.
Second, ask what the deliverable looks like at the end of the engagement. If the answer is a capability build, a trained team, a working workflow, or a campaign, you are in the capability or execution layer. If the answer is a strategic decision the partner stands behind, you are in the decision layer. The structural test is whether the partner's name is on the recommendation when it is challenged.
Third, ask what happens when the recommendation turns out to be wrong. A capability or execution partner will say "we will recalibrate the workflow." A decision-layer partner will say "we will tell you why the call was wrong, what we are changing, and what the new call is." The first is a service-level response. The second is an accountability response. The difference between the two is the difference between a consultancy that takes the work and a partner that takes the decision.
Cannes Lions begins June 22. The festival's theme is "The AI Hype Era Is Over, Proof Is the New Flex." According to AdPulse, that framing is the industry's first post-hype correction moment — a recognition that the buyer is tired of capability pitches and is asking for outcomes. The next two weeks will bring a flood of "AI consultancy" announcements: launches, partnerships, capability statements, workshop offers, embedded-team proposals. Most of them will be excellent capability or execution layer offerings. Almost none of them will answer the question that actually determines whether the buyer's AI investments produce a return.
The test for any "AI consultancy" pitch this quarter is whether the firm accepting the engagement is also accepting accountability for the strategic decisions the engagement is supposed to inform. If the answer is yes, the buyer has found a partner. If the answer is no, the buyer is buying capability or execution — and the strategic decision layer remains unoccupied, just as it was before the engagement began.
The strategic accountability layer is the one an AI-native strategy agency operates in. Not a tool. Not a capability training program. Not an embedded production team. A team that owns the strategic recommendation, defends it in the room where the decision is made, and recalibrates when the market moves. The synthesis cost of the six to twelve tools in your stack is absorbed, not pushed back to your team. The accountability is not an add-on. It is the product. One subscription. Decisions absorbed from the sprawl. End the sprawl. Get outcomes.
Ready to move from AI consultancy pitches to AI accountability? Get started with Autostrat.
Book a 30-minute demo. Bring a live question and watch the answer get built.