Start a project

Home / Insights / Article

Publicis, WPP and Ogilvy are rebuilding around AI. The playbook for everyone else

How Publicis, WPP and Ogilvy are putting AI into data, media and search, and what independent agencies and brands can borrow without a billion-euro budget.

The three biggest names in advertising are making a similar bet: the agency of the next decade is a data and technology company with creative talent attached. You do not need their budgets to learn from how they are doing it.

Publicis: AI under the hood

Publicis Groupe has taken the quietest route of the big networks. Rather than selling a flashy branded tool, it describes an AI layer, CoreAI, that sits underneath the whole group. According to the company's 2025 registration document, CoreAI connects the group's data, including 2.3 billion consumer profiles and trillions of campaign performance data points, and was built in-house by Publicis Sapient starting in 2023. A separate platform, Marcel, connects employees, and clients and partners have been able to join it since 2024.

The commitment behind it is large. In an interview with The Drum, CEO Arthur Sadoun said the group has been investing €12 billion over ten years in data, technology and AI, and that results were now showing up in financial numbers. eMarketer, summarising Publicis's full-year 2025 results, reports that the company says 73% of its operations and 80% of its media revenue are now AI-powered.

The same eMarketer piece adds a caution worth keeping in mind. Investors are asking whether the spend will pay off, and an industry survey it cites found nearly half of agencies, 46.4%, do not yet measure the return on their AI investment.

WPP: platform plus partnerships

WPP has leaned on its Open platform, which combines data, creative and production tools and uses virtual agents to simulate scenarios for brand strategy and audience planning. Ogilvy Switzerland, for example, has described running entirely on WPP Open.

WPP is also working with Google. Adweek reports that WPP-owned agencies are getting access to tools that help brands appear in AI search, including Google's AI Overviews, ChatGPT and Perplexity, as part of a wider partnership. One product, Generative Store, adapts a brand's website to each shopper's query.

Ogilvy: reputation is now a search signal

Ogilvy's Australia and New Zealand business launched Generative Impact, an offer built around the idea that a brand's reputation now decides how it is surfaced and cited in AI answers. It combines two things: earned media and PR work, led by Ogilvy PR, and the AI optimisation of a client's website, led by Ogilvy One. It cites Pew Research Center data showing people click traditional links less often when an AI summary appears.

That pairing is the interesting part. AI search engines draw on what is said about a brand across the web, not only what the brand says about itself. PR and SEO have always been separate budgets. They are starting to converge.

What the three have in common

Strip away the branding and the pattern is consistent:

  • Data comes first. Publicis's advantage is unified data, not a clever model.
  • Measurement is the open question. Even the largest groups are being asked to prove returns.
  • Search is being rebuilt. WPP and Ogilvy are both putting resources into visibility inside AI answers.
  • People stay in the loop. Publicis frames its approach as people first, with AI as an accelerant.

What independents and brands can borrow

A mid-sized agency or an in-house team cannot match €12 billion. The ideas still scale down.

Unify what you already own. Connect ad platforms, your website, your CRM and your sales data in one place, such as a warehouse like BigQuery. You do not need 2.3 billion profiles. You need your own customers in one clean view.

Build agents around one workflow at a time. Reporting, lead response and creative variant production are good starting points because the output is easy to check.

Treat reputation as part of search. Earn mentions on credible sites, keep your facts consistent everywhere, and publish clear, quotable answers to the questions your buyers ask. This is the work behind answer engine optimisation.

Set the measurement rule before you start. Decide what a good result looks like and compare it to a baseline. The industry's own surveys show this is where most teams are still weak.

The big networks have scale. Smaller teams have speed, and speed matters when platforms change every quarter. That is the gap an AI-native agency is built to use.

Sources

What are you
going to build?

Start a project