What was the main theme of Cannes Lions 2026?

AI Search and the agentic web dominated Cannes Lions 2026. The conversation shifted from creative execution to brand infrastructure: how brands get discovered and recommended by AI systems, and how publishers monetize content that AI engines cite but no longer send traffic to.

What is the agentic-web economy?

The agentic-web economy is the value chain in which AI engines read, synthesize, and cite content to answer users directly, and in which brand visibility and publisher revenue are tied to being inside the answer rather than ranked above it. It is the successor to the open-web link economy, where value moves from click to citation.

Why are publishers still important to AI if they are losing traffic?

AI engines rely on trusted publisher content to train and ground their answers, so authoritative publishers are cited more, not less. The break is in monetization: citation drives the answer but not the click, so traditional ad and affiliate revenue falls even as influence holds. Licensing helps the largest players but is not yet a market.

How is AI Search advertising different from SEO and SEA?

SEO and SEA optimize for ranking and clicks on a results page. AI Search advertising places brands inside the synthesized answer itself, in formats native to AI engines, and compensates the publishers whose content is cited. The unit of value shifts from click to citation, and SEO foundations like clean structure and authority still feed it.

Cannes Lions 2026 Highlights: AI Search Takes Over

Twenty years of Cannes argued about the ad. This year argued about whether brands and publishers even appear inside the AI answer. My highlights, decoded.

I have worked the Croisette for the better part of two decades, and Cannes Lions 2026 was the first year the festival stopped arguing about the ad and started arguing about whether brands and publishers will even appear inside the answer. AI Search was no longer a panel topic; it was the weather. These are the moments that stayed with me, and what twenty-five years in search tells me they actually mean.

This year's defining anxiety wasn't creative quality. It was visibility: will an AI engine surface your brand, and will it pay the publisher it quoted?
Cloudflare data presented at Cannes confirmed bots now outnumber humans online, and the click economy that funded the open web is structurally breaking.
Publishers are not obsolete but essential: AI needs their content as training and grounding fuel, yet pays a fraction of what traffic once did.
For brands, the new battleground is brand coherence, because the model reads every contradiction in your messaging back to the customer in real time.
The missing layer is a payment rail: an advertising model native to AI answers that compensates the publishers being cited. That is the category forming now.

What Cannes 2026 was actually about

Strip away the rosé and the beach cabanas and one shift ran through every serious conversation: the move from the open-web search economy to the agentic-web economy. The agentic-web economy is the emerging value chain in which AI engines read, synthesize, and cite content to answer users directly, and in which brand visibility and publisher revenue depend on being inside the answer rather than ranked above it.

The old Croisette and the new one

The old Croisette ran on the campaign, the impression, the click, and the traffic-funded publisher. The new one runs on the citation, the synthesized answer, the agent, and the publisher who gets quoted but not visited. The mechanic changed; the underlying needs did not. Brands still need to be seen. Publishers still need to be paid for being the source.

Matthew Prince and the moment the click economy broke

Cloudflare CEO Matthew Prince delivered the bluntest data of the week at Axios House on June 23: bots now outnumber people online, and humans increasingly trust AI answers without clicking the footnotes. He warned that AI's ability to compare everything with no brand loyalty could "destroy small businesses" and trigger mass consolidation.

On verification: Prince told the room that earning a click from Google is roughly ten times harder than four years ago, and far harder from AI engines. Cloudflare's own published crawl-to-referral ratios support the direction, running into the low thousands to one for OpenAI and the tens of thousands to one for Anthropic depending on the period, and Cloudflare Radar recorded bots overtaking humans at 57.5% to 42.5% in June 2026. The exact multiples move monthly; the structural break does not.

The most clarifying counterpoint came from People Inc CEO Neil Vogel, who described his AI licensing deals in restaurant terms: "all-you-can-eat" arrangements with OpenAI and Meta, an "à la carte" deal with Microsoft, and renewals expected on each. For years media feared AI would make content worthless. Cannes 2026 flipped that fear: AI needs trusted content more than anyone, and is willing to pay for it.

Scarcity is key. AI needs a model, power and inputs. We are the inputs.
— Neil Vogel, CEO, People Inc

Here is where I part ways with the optimists. The steelman is real: marquee publishers are signing eight-figure deals and AI is the best customer media has had in a decade. But licensing is lumpy, bilateral, and tilted toward the few brands big enough to be at the table. Vogel himself conceded People Inc cannot block Google, because its search and AI crawlers are fused, which he called an abuse of market power. Leverage without a payment rail is not a business model. It is influence stripped of recurring revenue.

The CMO 'how' problem and brand coherence

On the demand side, the panic everyone labeled an "SEO problem" was, as Digiday sharply put it on June 24, really about brand coherence finally getting graded in public. An LLM cannot be bought to the top of a page; it synthesizes whatever exists, contradictions included, and hands the customer one version of the truth. Search forgave a messy brand. The machine does not.

The numbers underline the gap: Boston Consulting Group surveyed 300 CMOs and found 96% believe AI is transforming marketing, but only about a third have rebuilt their workflows, operating models, and talent to capture the value. Marketers now compete for citations, trust signals, and machine-readable authority, which means brands are quietly writing two versions of themselves, one persuasive to people and one legible to models.

The human counterweight came from The Home Edit's Clea Shearer, who bought her brand back from private equity to protect creative control and refuses to let AI touch her content.

I want to believe things are real, and that's what AI lacks.
— Clea Shearer, co-founder, The Home Edit

What I'd tell brands, and what I'd tell publishers

For CMOs, media buyers, and agencies: ringfence the AI-visibility line

Treat AI visibility as a discipline distinct from SEO and SEA, with its own owner, budget, and KPIs. Audit how much of your funnel already runs through AI answers, fix brand coherence across PR, product, reviews, and content before chasing volatile "AI rankings," and ringfence a 2026 test budget for emerging AI Search ad inventory. Agencies should build this as a service line before a client asks why a competitor is being cited and they are not.

For publishers: monetize the citation, not the lost click

Stop pricing your future on traffic recovery and price it on being the source AI engines cite. The Press Gazette News Yacht showed the early playbook: Time is commercialising content designed to be read by AI agents, The Washington Post is selling sponsored answers in its own chatbot, and Hearst UK is monetizing its audience data. The revenue model that fits this role is citation-based advertising, with publishers inside the value chain rather than outside it.

The signals on the beach: agentic ad infrastructure

The forward-looking story was the plumbing. WPP Media launched a video Buyer Agent plus an agentic standards initiative with Disney, Netflix, NBCUniversal, Paramount, the IAB Tech Lab, and Prebid; Magnite introduced agent-to-agent orchestration; and ChatGPT ads landed in the UK. The industry is racing to build the rails on which AI agents transact. The unresolved question, the one I left Cannes thinking about, is which of those rails will actually pay the publishers whose content makes the answers worth trusting.

Conclusion

Hold on to this: Cannes 2026 made it official that the open-web click economy is giving way to an agentic-web citation economy, and neither brands nor publishers yet have a payment rail built for it. That rail is the category Smalk AI occupies, an AI Search ad network that places native ads inside AI answers while opening a new revenue stream for the publishers whose content fuels those answers. What I will watch next: which standards body or coalition sets the first pricing benchmark for sponsored citations, because that is the moment this market prices in.

My Cannes Lions 2026 Highlights: AI Search Ate the Croisette