BUSINESSTECH

|

|

5 min read

5 min

Omnicom Hands Off the Staff That Built Its AI Platform

Omnicom moved 468 people who built its Omni AI moat to an outside contractor. The reorg raises a hard question about how much of agency AI is owned and how much is rented.

By

Giovana B.

Omnicom has spent years telling clients that its Omni platform is a proprietary advantage, an artificial-intelligence moat that competitors cannot easily copy. An Adweek investigation now reports that hundreds of the people who built that moat no longer work for Omnicom. In June and July the holding company transferred at least 468 staffers from Omni’s product and engineering teams to a third-party contractor, Endava, and laid off roughly 50 more. The move complicates the story a company tells when it sells its technology as the reason to hire it.

What Happened

The transferred employees span the United States, the United Kingdom, India, and Malaysia, and most held product and engineering roles, the core technical talent behind the platform. According to the reporting, Omnicom cut about 50 US workers from the Omni Platforms Product and Engineering division on June 9, then routed the larger group to Endava over the following weeks. Internal accounts describe a reorganization that created uncertainty around roles, reporting lines, client communication, and day-to-day operations. Whatever its long-term logic, the change moved a substantial share of the people who understood the platform outside the company that markets it.

The Moat Question

Omni is Omnicom’s flagship data and AI system, the thing the holding company points to when it argues that it can target, plan, and measure better than a rival or a consultancy. A moat implies ownership, control, and a capability that lives inside the walls. When the engineers who built and maintain that capability become contractors at another firm, the claim gets harder to defend. Outsourced talent can still deliver good work, but it changes the nature of the asset from something the company owns to something it procures, and procurement is exactly what a moat is supposed to prevent. The gap between calling a platform proprietary and staffing it through a vendor is the gap clients will now probe.

The Cost Logic

The decision does not come from nowhere. Omnicom is integrating its merger with Interpublic and has doubled its cost-savings target to 1.5 billion dollars, with thousands of jobs across the combined company under review. Shifting engineering work to a lower-cost contractor is a recognizable margin move, the kind of restructuring holding companies reach for when they are under pressure from investors, from consultancies encroaching on their territory, and from clients pulling work in-house. Read through that lens, the Endava transfer is less a statement about AI strategy than a line item in a savings plan. The problem is that the savings plan collides with the marketing story.

The Contradiction the Industry Keeps Hitting

Agencies are selling AI transformation to their clients at the same moment many are thinning their own technical ranks, and the two messages are hard to reconcile. A holding company that pitches itself as an AI leader while offloading the team behind its signature platform invites the question of whether its AI is a capability or a narrative. Continuity is the immediate risk. Clients who built campaigns and data pipelines on Omni will want to know who maintains the system, who answers when it breaks, and whether the roadmap they were sold survives the reorganization. Trust in agency technology depends on the sense that someone accountable owns it, and a messy handoff erodes that sense.

What It Signals for the Industry

The episode is a window into the strain on the holding-company model. AI is simultaneously the industry’s most promoted promise and one of its heaviest cost centers, and the tension between those roles is forcing awkward choices. Some agencies will treat proprietary technology as a genuine core competency and invest to own it outright. Others will conclude that building and maintaining large platforms is a commodity best rented, and will compete on services and relationships instead. Omnicom’s move suggests which way the cost pressure pushes, at least for now. For an industry that has staked its relevance on owning the machinery of modern marketing, outsourcing that machinery is a consequential tell.

What Marketers Should Take

For marketers choosing partners, the lesson is to look past the platform demo and ask who actually builds and runs the technology behind it. An impressive AI story is not the same as an owned AI capability, and the difference shows up when a system needs to evolve or a problem needs an owner. Clients should ask where the engineers sit, how continuity is guaranteed through a reorganization, and what happens to their data and their roadmap if the vendor relationship changes. None of this means outsourced technology cannot perform. It means the marketing claim and the operating reality should match, and the burden is on the agency to prove that the moat it advertises is one it still holds.

To access this article, become a WAM member. Subscribe

ONE MEMBERSHIP • EVERYTHING INCLUDED

Every story that moves marketing.

And an AI that reads them for you.

Powered by 1000+ WAM insights, WAMer turns your research into answers in seconds – summarized and sourced.

1 credit = 1 question asked to WAMer. Credits refresh every month and unused credits don’t roll over. Your trial starts with 5 credits; the full monthly credits unlock when your plan begins.

START HERE

3-Day

Free Trial

No charges for 3 days

Renews at

$12.00

/month

Cancel anytime

OR SAVE 55%

Quaterly

$6.67

Per month

Billed

$20.00

every 3 months

Cancel anytime

FURTHER READING