Emma Labrador, VP Marketing at Displayce, shares her key takeaways from Cannes Lions 2026 and explains why the rise of agentic AI will reshape the way media decisions are made...

Agentic AI was already a major topic before Cannes. On the Croisette, it became impossible to ignore. In meetings with agencies, brands and technology partners, the same question kept surfacing: how far can

AI move from assisting work to actively shaping decisions and actions?

That shift was visible across the festival. Cannes Lions chairman Phil Thomas said that 40% of entries used AI in some form this year, twice the 2025 share. But for me, the most significant change was not simply the growing use of AI. It was the emergence of agentic models, where AI interprets, compares, recommends and acts across the decision process. Three points stood out:

1. Agentic AI was everywhere. But what actually makes a good agent?

The term “agent” is now being applied to a growing range of tools, assistants and interfaces. Across the industry, initiatives are multiplying, from WPP’s Agent Hub and the IAB Tech Lab’s work on agentic advertising standards to a wave of new agentic planning, buying and optimisation tools showcased across Cannes. But behind the excitement, a more useful question is emerging: what actually makes a good AI agent?

An agent does not become valuable simply because it can answer a question, produce content or automate a task. Its value depends on the intelligence behind it: the quality of the data it can access, the depth of its business expertise, the tools it can use and the context it has to support a decision.

In media planning, that means going beyond an AI assistant producing a generic recommendation. A useful agent should be able to interpret a brief, explore relevant data, compare scenarios and explain why one option may be better suited than another.

For DOOH, this requires highly specialised intelligence. At Displayce, this foundation has been built over 12 years of expertise in the media, supported by a data ecosystem spanning audience, mobility, inventory and contextual signals, as well as exclusive JCDecaux data. This depth of knowledge is what allows AI to move beyond generic recommendations and reflect how DOOH decisions are actually made.

The channels that can be properly discovered, understood and evaluated at that point will naturally have a better chance of entering the plan. For DOOH, this is a major opportunity. The medium still enters many planning processes too late, once budgets and channel choices have largely been decided. Bringing DOOH expertise and data into the AI tools used by media teams could help the channel be considered earlier. It could also provide a clearer rationale for the role DOOH can play in the media plan.

This is what we mean by Agentic DOOH: making specialist DOOH intelligence, data and activation capabilities accessible to AI systems, so the media can be considered earlier and more effectively within the decision process. This thinking shaped our own announcement at Cannes, where Displayce introduced three specialised agents accessible via MCP, supporting planning, activation and analysis. 

2. The agentic shift is also a battle of protocols

Another recurring topic at Cannes was MCP (Model Context Protocol) and AdCP (Agentic Campaign Protocol), two different approaches to how agents interact with the advertising ecosystem.

Put simply, MCP allows an AI agent to connect to and operate existing tools, data and platforms. AdCP explores how advertising agents could communicate directly with one another across tasks such as planning, buying and measurement.

For us, the immediate opportunity lies in MCP. This is why Displayce chose to make its specialised agents accessible through it: to bring DOOH expertise and capabilities into the AI environments media teams are already using, while keeping human judgement and validation at the centre of the decision process.

3. Faster decisions make measurement even more important

Across our Cannes meetings, measurement remained one of the most consistent concerns. Brands and agencies want clearer proof of incremental reach, brand impact, store visits and sales uplift. They also want to understand how DOOH contributes alongside other channels. If more recommendations are supported by AI, teams need to understand why a channel, audience or context was selected and whether the decision delivered the expected outcome.

A recent McKinsey post-Cannes analysis captured this gap well: nearly 60% of marketers use AI several times a week, yet only 10% have fundamentally redesigned their workflows to capture meaningful business impact.

That is the challenge ahead. Faster recommendations are only useful if the underlying signals are reliable and the outcome can be evaluated. 

Measurement therefore needs to feed the next decision. Which contexts performed best? Did exposure influence footfall, brand consideration or sales? What should change in the next plan?

This is another area where agents can add real value. By bringing together data from brand lift, drive-to-store and sales lift studies, they can simplify reporting, make results easier to interpret and help teams demonstrate the contribution of DOOH more clearly.

What do we take away from Cannes 2026?

Agentic AI is becoming part of how advertising work gets done. The biggest change may happen before a campaign is activated, at the point where a brief becomes a recommendation and budgets start to move.

For DOOH, this is a chance to solve a long-standing problem: making the medium easier to consider early, easier to justify and easier to connect with the wider media plan. That, ultimately, is the opportunity behind Agentic DOOH: not simply adding AI to existing workflows, but making specialist DOOH intelligence available at the moment media decisions are being shaped.