Cannes Lions 2026: when AI can make anything, taste becomes the differentiator

You could file Cannes Lions under "one for the agency crowd" - a week of rosé on the Croisette, awards trophies, and marketers congratulating each other in the south of France.

Here's why that filing is wrong.

This year's Festival drew on more than 150 hours of stage content across five streams, and nearly every session bent toward the same question now sitting on your desk: when AI can generate anything, what actually creates value? That's not a marketing question. It's a hiring question, a budget question, and a strategy question. Our team covered the full run as accredited press, Sunday through Friday. Here's the signal, stripped of the spritz.

Last year the Palais staged a defence of humanity against the machines. This year the framing flipped: AI was treated less as an existential threat and more as plumbing - useful, unavoidable, and no longer the story in itself. Instagram's Tessa Lyons put the optimistic case plainly, calling AI not the end of human creativity but a new chapter that amplifies it. But the more consequential point ran underneath the optimism. As production barriers collapse and everyone holds the same tools, the scarce asset stops being output and becomes judgment - the taste to know what's worth making. Publicis Groupe CEO Arthur Sadoun described a revival of senior people who can tell great work from poor work. The juries backed it up: fully AI-generated work consistently fell flat when human creative direction was missing.

For an operator, this moves the money. The bottleneck is no longer production capacity, which is now cheap - it's editorial judgment, which isn't. As Unilever's Leandro Barreto warned, the industry has confused motion with meaning and become efficient at making things people ignore. The move this quarter: stop rewarding volume, and start paying for the people who can reliably tell good from good-enough - then put them close to the decisions.

The sharpest strategic idea of the Festival came from the Creative Impact stream: the shift from consistency to coherence. Effectiveness expert Les Binet reframed coherence as all the things a brand does adding up to a whole - not repeating the same asset everywhere, but making sure every touchpoint, however different, belongs to the same story. Klarna's CMO David Sandstrom put the practical version bluntly, describing brand work at scale as less about creativity and more about infrastructure. The reason it matters now: brands increasingly can't control the channel. Sephora's sponsorship principle was "never do a logo placement," and its sports investment still returned an average of 12x - because its coherence travelled into someone else's moment without a badge bolted on.

The translation for leaders: in a landscape with more channels and more judges than any brand can directly influence, control moves from owning each touchpoint to building a brand coherent enough to show up recognisably across a fandom you don't own, a creator you don't script, or a model you can't predict. That's a systems investment, not a campaign line - and it sits with whoever owns your infrastructure, not just your marketing.

The most operationally useful research at Cannes came from INSEAD, Jellyfish and WARC: almost no correlation between how humans and large language models judge the same creative work. P&G's Marc Pritchard framed the consequence - brands must now build memory in people's minds and algorithms at the same time. Jellyfish's Tom Roach argued the new bar for effectiveness is work that survives ranking, retrieval and recommendation by machines as well as moving people. The reassuring part, for brands that have done the hard work: models reward exactly the kind of distinct, consistently expressed brand platform that good brand-building already produces.

That gap - between how you show up to humans and how you show up to the machine now mediating discovery - is the new front line. Being legible to models is a growth channel, not an IT detail. The test is cheap and worth running this quarter: ask a mainstream AI assistant to find, describe and compare your brand, and watch what it gets wrong.

The counter-signal ran through every stream: the era of winning on scale is closing. AI's pattern-matching threatens what Orchard's David Kolbusz called a tsunami of the mediocre, the safe and derivative. The strategic response wasn't to automate harder but to go the other way - toward friction, distinctiveness, and community. Patagonia's Alex Weller made the trade explicit: the more clearly you state your values, the less universally acceptable you become, and a watered-down message is more dangerous than a polarising one.

Two things for operators to clock. First, creators moved from garnish to embedded operators - Mastercard's Kaveri Khullar's phrase - because meaning increasingly travels through people a brand doesn't control, and memory, not reach, is what compounds. Second, the scarce resource underneath all of it is attention and trust, both of which now form at community level. As Reddit's Laura Nestler put it, brands have to act as participating citizens, not billboards. In an infinite-content world, the question to put to your own team is which communities you actually participate in, and which you only advertise at.

The line that stuck to the week wasn't from a keynote stage - it was Rory Sutherland's aside that nothing scales better than mediocrity. Delivered as a joke, it worked as a diagnosis. Scale was the thing everyone spent a decade optimising for, and it turns out to be the one quality AI hands out for free. Whatever you were winning on because you could do it bigger than the next firm, you're now winning on by a shrinking margin.

Edelman's data put a number under the mood: a large share of consumers are now insular, trusting only friends, family and like-minded others - and for that group, earned attention is roughly five times more powerful than paid. That's the uncomfortable half of the Festival. Reach is getting cheaper and less useful at the same time, while the things that don't scale - being trusted, being specific, being in the room - hold their value. Worth clocking: the budget line that looks most efficient on a spreadsheet is usually the one buying you the most mediocrity.

That's the read from Cannes Lions 2026. Beneath the awards and the activations, the through-line was consistent - as AI absorbs the doing, the distinctly human work of judgment, taste and connection becomes the differentiator, not the overhead. Want the primary material? LIONS has published the Official Cannes Lions Wrap-Up 2026, which draws on the full 150 hours of stage content across the five streams and is worth a look if you want the source behind any of the above.

We covered Cannes Lions as press delegates for the full run, spending the week on the Croisette with founders, brand leaders and investors - a lot of those conversations are being lined up for our AI Central Voices interview series over the coming weeks.

If any of the above is live for you, two conversations we're always glad to have:

Brands and partners - if you're an AI or SaaS company trying to reach 300,000+ senior operators, founders and C-level readers, we can talk about how a partnership with AI Central turns that attention into pipeline.

Consultancy and advisory - if you're a founder or operator trying to separate the announcements from what actually changes your next two quarters, that's the conversation we most enjoy.

Email us: [email protected]

More to come - founder interviews are being edited now, and we'll have follow-ups on the judgment-versus-production shift and the machine-readable brand question specifically. If there's a thread from the week you want us to pull, let us know.

AI Central reaches 300,000+ professionals across LinkedIn and newsletter. Forward this to whoever on your team is still measuring output instead of judgment.