Omnicom Quietly Transfers 468 Omni AI Engineers to Endava — After Selling the Platform as Its Big AI Bet
Updated: Sep 1
TL;DR: Adweek documents Omnicom transferred at least 468 Omni AI engineers and product staff to contractor Endava this summer while continuing to pitch Omni as its central AI capability to clients.
What Adweek uncovered
Documentation obtained by Adweek and published on 20 August 2026 shows Omnicom transferred at least 468 product and engineering staff who built and maintained the Omni AI platform to Endava between June and July 2026. Roles were affected in the US, UK, India and Malaysia. A further ~50 US employees on the Omni Platforms Product & Engineering team were laid off on 9 June 2026.
The move is striking because Omnicom CEO John Wren specifically highlighted the refreshed Omni platform as a centrepiece of the group's client offering at CES in January. In official disclosures and investor updates, Omni has been positioned as the AI 'nerve system' that will justify Omnicom's headline valuation and its case for merging with Interpublic Group.
Why offload the very team you're selling
There is a plausible defensive read: Endava is a specialist engineering partner, offloading gives Omnicom variable rather than fixed cost, and the platform's IP still lives with Omnicom. Big holdcos have long outsourced parts of their martech build stack — this could be Omni entering a mature-operations phase.
The less flattering read is that Omnicom is trying to lift near-term margins ahead of the IPG combination close, and needs to show cost synergies fast. Offloading a US-heavy engineering base to Endava — publicly a lower-cost near-shore/off-shore provider — is one of the fastest levers available.
Marketing Minute's read
US CMOs who buy media through Omnicom should ask three specific questions in Q4 reviews. Who now owns the Omni product roadmap — Endava or Omnicom? What are the SLAs on new-feature releases now that a contractor sits between planners and code? And how does data ownership work when a third party operates the pipes? The answers matter, because Omni is often sold as the reason for full-service fees over point-solution vendors.
For the broader industry, this joins WPP's Open X reshuffle and Publicis's LiveRamp acquisition in a summer that has quietly repriced 'holdco AI'. What clients thought was proprietary is increasingly a stitched-together stack of partners — and that has implications for the price of the strategic layer sitting on top of it.
Quick FAQ
Q: What did Omnicom do?
A: It transferred at least 468 engineering and product staff who built its Omni AI platform to third-party contractor Endava between June and July 2026, with a further ~50 US layoffs on 9 June.
Q: Why does this matter?
A: Because Omnicom has publicly pitched Omni as its central AI capability and a key reason clients should stick with the holdco — offloading the team behind it raises questions about how strategic Omni really is inside the group.
Q: What should Omnicom clients ask?
A: Who owns the product roadmap now, what SLAs govern new-feature releases through Endava, and how client data flows are governed with a contractor operating the pipes.
When you publicly pitch a platform as your AI moat, then quietly hand its engineers to a contractor, it's rarely just cost-cutting — it's a signal that the pitch and the P&L don't yet agree. Every US client on an Omni deck this fall will and should ask why. — Daniel Nikolla, Founder of Merx Marketing
The takeaway: US marketers should treat 'holdco AI' as a stitched-together stack now, not proprietary IP — and price the strategic layer accordingly.
How much of your agency's AI 'proprietary tech' is actually a contractor with a nice logo? Ask this quarter, not next.
If you would like to know more about this topic, please contact us on danieln@merxmarketing.co.uk
Sources: Adweek exclusive (Kendra Barnett), Storyboard18, TVNewsCheck, Endava investor filings.
Related reading
Written by Daniel Nikolla, Founder of Merx Marketing Ltd and Marketing Minute
Comments