While most agencies are still showcasing their AI party tricks, Publicis just showed their hand. And it's very good.
Last week, with little fanfare but enormous implications, Publicis Groupe unveiled its acquisition of Lotame – the world's leading independent data and identity solution. This strategic move isn't just another holding company acquisition; it's a fundamental redefinition of what AI in marketing can and should be – backed by a $1.5 billion investment in acquisitions over the past six months to build out their CoreAI vision.
The timing couldn't be more revealing, showing the stark differences in the major players. In the same week WPP announced tighter integration with Stability AI, essentially giving their staff better access to image generators (something that is already table stakes for months), Publicis demonstrated they're playing at an entirely different level. Actually, a different game.
Beyond "Will Smith Eating Spaghetti"
If you've been in any marketing presentation in the last 18 months, you've seen the obligatory GenAI demos – the text-to-video wonders producing "Will Smith eating spaghetti" or AI-crafted copy that sounds suspiciously like every other AI-crafted copy. These parlour tricks initially dazzled us but ultimately proved to be the shallowest expression of AI's potential in marketing.
Publicis has moved well beyond this. With the Lotame acquisition, they've revealed what should have been obvious all along: the true power of AI in marketing isn't in content generation alone – it's in the connection between identity resolution and personalised content delivery at scale.
This isn't just theory. They've already launched 50 pilots of CoreAI demonstrating the real-world impact:
A financial services client using probabilistic modeling to identify consumers likely to become high net worth by analyzing historical patterns of those who did:
CPG advertisers assessing how inflation impacts consumer decisions to trade up or down between premium and value brands: and
Retail media optimisation that dynamically adjusts media bidding based on local store conditions like stockouts.
Identity: The Missing Piece
The numbers from the acquisition are staggering: Lotame brings 1.6 billion unique IDs built from 100+ data sources across 109 countries. Combined with Epsilon's existing database, Publicis will now have profiles on nearly 4 billion individuals, covering over 90% of global adult internet users.
But raw data isn't the breakthrough. It's what Publicis plans to do with it. By combining Lotame's identity solution into the existing Epsilon-powered CoreID identity solution and then with their CoreAI platform, Publicis isn't just creating bigger audience segments – they're building a system that can recognise the same consumer across devices and platforms and deliver precisely the right message at the right moment.
As Arthur Sadoun, Publicis' CEO, bluntly put it: "In the age of AI, the name of the game is connect or die."
The Strategic Moats in Action
I've written previously about the strategic moats that Publicis has built over the years, and this move demonstrates exactly why those moats matter:
Media Scale: With this expanded identity graph, Publicis can deliver more precise targeting across their media operations, strengthening their already formidable buying power. While others turn their attention to principal media buying, Publicis steps beyond this.
Data at the Core: By integrating Lotame with Epsilon, they've expanded their first-party data moat into global territories.
Technology Prowess: This isn't just about data collection – it's about activating that data through their CORE AI engine to drive personalised experiences. The 20,000+ engineers that came with and are built out from the Sapient operation mean they are building their own technologies that they can charge for, not using off-the-shelf reskins like others;
Organisational Culture: The most important of the moats. With their "power of one" approach, Publicis continues to operate as a unified company, with this technology available across their entire network. It will be adopted fasted, learned faster, applied faster and improved faster because of this.
Beyond Audience Segmentation
What makes this approach particularly powerful is that it goes beyond traditional audience segmentation. While competitors focus on using AI to create content or optimise media spend in isolation, Publicis is building an end-to-end system that connects identity to content creation to media activation – all within a closed loop.
This means they can potentially recognise an individual consumer, generate personalised creative specifically for them, deliver it across the right channels, measure the response, and continue refining – all within an integrated system. It's the marketing equivalent of a self-improving flywheel, where each interaction makes the next one more effective.
The Category of One: Beyond Holding Companies
When Sadoun claims Publicis has created a "category of one," it's not just corporate bravado. There's truth in the statement. No other holding company has assembled this specific combination of assets and capabilities. WPP has strong creative agencies and media scale but they lack integration, they have no first-party data solution and lack real technical capability at scale (It's reported the The Coca-Cola Company is considering between the incumbent WPP and Publicis for their North American media business, so let's see…) Omnicom has made strides with its Omni platform but doesn't have the identity scale - which is the main attraction of their IPG purchase which will bring them Axciom, but even a perfect integration of this will be too late in my view.
But the true competition isn't just other holding companies. It's the major forces that are attempting to build similar AI-powered marketing ecosystems.
Accenture and other consultancies have been aggressively building martech ecosystems that clients can outsource to on a build-operate-transfer model. They leverage their enterprise integration expertise and C-suite relationships to capture increasing portions of marketing budgets.
Meanwhile, the major platforms (Google, Meta, Amazon) continue expanding their direct-to-client toolsets, attempting to disintermediate agencies entirely. They leverage their vast first-party data and closed ecosystems to offer increasingly sophisticated targeting and measurement.
In addition to making the obvious claim of agnosticism, and stoking the fear that clients have of putting too many eggs in the platforms baskets, Publicis appears to be countering this by implementing publisher-friendly strategies – integrating Lotame with Epsilon's first-party tagging system to help publishers navigate cookie deprecation, creating more robust publisher networks that benefit from their identity solution. This could create a more open alternative to the walled garden approach of platforms.
The Retail Media Revolution
One of the most profound implications of Publicis' approach is in the rapidly growing retail media space. With retail media networks becoming increasingly important (Amazon Ads, Walmart Connect, Kroger, etc.), Publicis' enhanced identity capabilities could be transformative.
Their private Demand-Side Platform (DSP), powered by this expanded identity graph, enables several critical advantages in retail media:
Cross-retailer attribution: Tracking consumer journeys across multiple retail environments, something individual retail media networks can't offer. This addresses one of retail media's biggest challenges: fragmentation across multiple networks, each with their own data silos and attribution models.
Enhanced ROAS measurement: Connecting ads seen on one platform to purchases made on another, closing the loop on attribution.
First-party data activation: Enabling brands to activate their own first-party data across retail media environments while preserving privacy and identity.
New Commercial Models: Escaping "The Great Compression"
Perhaps most revolutionary is how Publicis' CoreAI approach could fundamentally transform agency commercial models. As I've written about previously in "The Great Compression," the marketing services industry faces unprecedented pressure from all sides – clients demanding more for less, technology automating core services, and platforms disintermediating traditional roles.
Publicis' AI capabilities open the door to entirely new commercial approaches:
SaaS-style subscription models: Moving from billable hours to ongoing access to AI-powered marketing capabilities.
Outcome-based compensation: Tying agency compensation directly to measurable business results – from automotive test drives to e-commerce transactions – enabled by more accurate attribution and conversion tracking.
Variable pricing based on AI effectiveness: Commercial models that scale with the demonstrated return on marketing investment.
Privacy-first processing options: By developing capabilities to deploy AI models within client-owned data environments and creating a "probabilistic data workbench" for transparency in AI-driven projections, Publicis is building trust mechanisms that can command premium pricing.
By connecting identity resolution to business outcomes, Publicis has the potential to break free from the industry's long-established pricing models based on time and materials, or even outputs-based models. Instead, they can price based on the value they create – a fundamentally different approach that could help escape the compression affecting the rest of the industry.
And to Publicis, from me at least, 'Chapeau'.
Brandflow is written by Justin Billingsley, who has spent his career on all three sides of the industry's table: senior client, global agency leader, technology founder. First published 9 March 2025 in the Brandflow newsletter on LinkedIn.