What Cannes Lions 2026 Signaled About the Next Twelve Months of AI in Marketing

Cannes Lions 2026 was the year marketers started taking AI seriously.

Twelve months ago, the Croisette was full of ambition - big experiments, bigger promises. This year, the tone was noticeably more disciplined. The question on stage and at every dinner table was no longer what AI could do, but where it actually creates value - and, just as often, where it shouldn't be pointed at all.

If I had to compress the week into one line, it would be this: marketing leaders aren't afraid of agents - they're impatient for agents that respect the community layer instead of spamming it with senseless automation.

I attended this year as the founder of Blueberry, an AI marketing platform that turns social engagement into revenue. Over the course of the week, I drew insights from the people running marketing at L'Oréal, Strava, Victoria's Secret, Sephora, Levi's, Oura, VIZIO, Grindr, GoGo squeeZ, and more across dozens of panels.

I also got to speak on a panel alongside our incredible customers, such as CMO of the San Francisco 49ers, and technology partners like the CMO of Hootsuite; and had an opportunity to host marketing leaders across numerous dinners and cocktail receptions.

Out of all of it, a consistent picture emerged of what marketing leaders actually care about. Three trends stood out - and they point to specific predictions about what gets hot in the months ahead.

Trend 1: Taste is the new scarce asset

If there was one word that dominated the week, it was taste. As content generation becomes easier, execution is commoditizing - and the industry has a name for the failure mode: mediocrity at scale. When every brand runs the same models trained on the same patterns, output converges toward sameness. Volume stops being an edge when everyone has volume.

Asmita Dubey of L'Oréal framed it precisely on her panel: AI isn't replacing brand judgment, it's raising the bar for it. Offloading decisions to the machine loses the plot; the discipline is being deliberate about where you deploy it. The competitive question has shifted from "are you using AI" to "does your use of AI make you more distinctive, or just faster at being average."

The awards told the same story. Entries fell roughly 25% this year under tougher integrity criteria, forcing work to win on the idea rather than the entry engineering. The festival also introduced its first-ever Creative Brand Lion - awarded to AB InBev - to recognize the internal systems and culture that make creative excellence repeatable, effectively institutionalizing the view that judgment is an organizational capability, not a campaign. And in a year saturated with generated visuals, Apple TV's deliberately handmade rebrand took top design honors - taste as a stated position.

The organizational corollary came up repeatedly too: the leaders getting real value are not layering AI onto existing processes. They are redesigning workflows around it, with sponsorship from the very top. That is slower and harder than buying tools, which is exactly why it separates winners from the pack.

The brands that win with AI will not be the ones that automate the most. They will be the ones that know what not to automate.

Trend 2: Community is replacing impressions as the unit of marketing

The second thread ran through nearly every brand session I attended. Strava's growth isn't driven by mileage tracking but by belonging - Gen Z joins for the community and the cultural moment. Victoria's Secret let 100,000 applicants shape its fashion show through an open casting call. Sephora's partnership with the Valkyries works because it started as genuine interest, not a media plan. And Grindr's framing stayed with me: a brand is who you say you are, who people say you are, and who you actually are - kept in alignment, with trust and safety as the foundation. Fittingly, the Social & Creator Grand Prix went to Heineken for turning forwarded voice notes into free beers at real bars - a digital habit converted back into the oldest social behavior there is.

This is also where AI becomes both powerful and dangerous. You can automate tasks inside a community, but you cannot fake care. You can speed up a response, but you cannot replace judgment. Every interaction either compounds trust or erodes it - there isn't much neutral ground.

I had the chance to speak on a panel of my own alongside the CMOs of the San Francisco 49ers and Hootsuite, on where AI takes marketers next - and the questions from the room weren't about content generation. They were about community: how to automate without eroding the trust that took years to build.

The lesson was not that every brand needs a bigger community strategy. It was that community is no longer downstream of marketing. It is becoming the place where brand truth is tested.

Trend 3: Social engagement is becoming a revenue channel

Once that is true, social engagement stops being just engagement. It becomes a revenue channel.

Brands are done renting attention. At Cannes, the conversation was not about whether social matters - everyone agrees it does. It was about why so much buying intent still disappears inside replies, DMs, and comments instead of turning into anything a brand can act on. Communities remain one of the most under-monetized assets most brands have, not because the demand is not there, but because the infrastructure to capture and measure it has lagged behind.

At the dinners we hosted, what surprised me was how quickly conversations with CMOs moved from polite curiosity to practical urgency once we showed how Blueberry approaches this layer - turning social engagement into a measurable, revenue-driving channel rather than a support function. The appetite was not coming from marketers looking for another content tool. It was coming from marketers who had already concluded that their community was sitting on untapped revenue, without the infrastructure to prove it.

One way to make sense of all three trends is to separate the execution side of marketing from the relationship side. Production, adaptation, response routing, and workflow automation are where agents are already creating leverage - helping teams move faster and capture demand that would otherwise disappear.

But taste, trust, belonging, timing, and the decision of when a brand should speak at all are relationship problems, not just execution ones. They are the parts of marketing where people decide whether a brand feels real. The next phase of AI in marketing will not be defined by how much brands automate, but by whether they use automation to protect and deepen those relationships. Agents can support that layer. They do not get to own it by default.

What gets hot in the next few months

Based on what the leading brands are already doing, here is where I'd place my bets for the second half of 2026:

Agents move into critical marketing workflows. The shift from chat-style copilots to agents embedded in revenue-driving processes - social, lifecycle, commerce operations - is already underway inside the most sophisticated marketing orgs. Expect budget to follow, and expect "agent-ready" to become a standard question in every MarTech evaluation.

Social engagement becomes a revenue channel, not a support function. Expect reach to lose its seat as the primary social KPI, replaced by conversion metrics on replies, DMs, and comments. A concrete marker to hold me to: by mid-2027, leading consumer brands will report revenue attributed to social engagement, the same way they do for email and SMS.

First-party data from conversations becomes the prize. As content and marketing output becomes easier to produce, the edge shifts to striking a personal chord with every individual. Mature marketing organizations will start logging every interaction as a signal for their AI marketing agents. Each reply, question, and DM is a customer telling you what they want in their own words: first-party data richer than anything a pixel ever captured, arriving exactly as privacy changes make it more valuable than ever. Expect sophisticated orgs to stop treating conversations as ephemeral and start treating them as a data asset - piped into CDPs, feeding personalization engines, and informing product. The brands that capture and act on conversational data will out-personalize everyone else; the rest will keep buying increasingly blind third-party lookalikes.

Social AI with a trust layer. Community-facing automation will bifurcate: tools that generate volume into feeds will lose ground, while agents designed around brand safety, authenticity, and community health will win the enterprise. Trust and safety stops being a compliance line item and becomes brand infrastructure.

IRL and earned media re-rate upward. Precisely because generated content is commoditizing, the things that can't be - live experiences, genuine creator partnerships, earned coverage - will command a growing share of attention and spend. Expect the best brands to fund this reallocation with the efficiency gains from automation, not in spite of them.

The through line

A week surrounded by the most advanced technology in our industry left me more convinced of something almost contrarian: marketing is heading somewhere more human, not less. The winners of the next twelve months will automate relentlessly - and use every hour they get back to build taste, community, and trust that no model can replicate.

That's the bet we're making at Blueberry. Judging by the rooms at Cannes, we're not the only ones.

—

Nima Mozhgani is the founder of Blueberry, an AI marketing platform that turns social engagement into revenue and a portfolio company of UpScaleX. He attended the 2026 Cannes Lions International Festival of Creativity, where he spoke on a panel about AI and the future of marketing.