Skip to content

We Asked What AI Couldn’t Replace. Cannes Lions 2026 Gave Us The Answer.

A week before Cannes, we asked a simple question: if AI can generate almost everything, what becomes scarce?

Then Cannes happened, and the answer did not arrive as a keynote line. It arrived sideways, through the winning work, the new award architecture, the creator takeover, the quieter scepticism around AI, and the campaigns people kept mentioning after the panels ended.

The festival did not feel like a celebration of infinite generation. It felt like a correction.

For two years, marketing has been fascinated by output. More copy. More images. More edits. More variants. More speed. More ways to make the same calendar look full.

Cannes Lions 2026 seemed to ask a colder question: what if more was never the problem? What if the scarce resource was knowing what deserved to exist?

That is the second chapter of the Anti-Slop Marketing Stack. The first piece was a hypothesis. Cannes gave us evidence.

Contents

The Festival Answered A Different Question

Most people arrived expecting an AI festival. That would have made sense. Agentic AI was everywhere in the week’s vocabulary. Vendors arrived with demos. Platforms arrived with automation promises. Marketing teams arrived with the familiar mixture of hope, fear, and procurement curiosity.

The louder story looked obvious: AI would dominate Cannes.

The quieter story was more interesting. Human creativity took the room back.

Business Insider’s Cannes recap described a more grounded mood than last year. Leaders talked openly about AI’s limits. PMG CEO George Popstefanov said he was hearing more discussion of what AI cannot do. EY CMO John Rudaizky warned about the illusion of accuracy and completeness. Fernando Machado of Chipotle argued that the industry had become too obsessed with optimisation, dashboards, and shortcuts.

The tone had changed. Last year, the question was whether AI could make advertising. This year, the better question was where AI belongs inside creative organisations.

That distinction matters. A tool can generate. An organisation has to judge.

Discovery 1: AI Became Infrastructure

The most interesting thing about AI at Cannes was how ordinary it became. It was present everywhere. It also stopped being enough.

Nobody gives awards to electricity for powering the projector. Nobody praises a spreadsheet for doing arithmetic. Once a capability becomes expected, the prestige moves somewhere else.

AI is approaching that point in marketing. A model that generates a storyboard is useful. A model that writes fifty variants is useful. A model that resizes, translates, tags, summarises, and adapts assets is useful.

Useful no longer means distinctive.

The work Cannes rewarded tended to sit above the tool layer. It had context. It had timing. It had social intelligence. It had a reason to exist outside the dashboard.

This is the first signal from Cannes 2026:

AI became less interesting as a creative object and more important as creative infrastructure.

That is maturity. It is also a warning. If every team has the same generator, the generator cannot be the moat.

Discovery 2: Creativity Became An Organisational Capability

The most important Cannes shift may not be a campaign. It may be the new Creative Brand Lion.

The category signals a change in what the industry wants to reward. A single campaign can still win. A beautiful piece of work can still matter. But Cannes is also beginning to recognise the organisation that repeatedly produces exceptional work.

That is a different theory of creativity.

The older model looked like this:

Great idea
↓
Great campaign
↓
Award

The newer model looks more like this:

Culture
↓
Creative operations
↓
Brand memory
↓
AI and production workflows
↓
Evaluation
↓
Repeated creative excellence

This feels close to what software engineering learned years ago. At first, companies admired individual programmers. Then they learned to admire engineering systems: version control, testing, deployment, observability, code review, shared standards, incident learning.

Marketing is moving in the same direction. Great campaigns are outputs. Great creative systems are the product.

This is where AI changes the organisational question. A brand does not need an AI that makes more ads. It needs a creative operating system that helps teams remember, decide, produce, test, and learn.

Discovery 3: The Winners Were Useful

The strongest winning work did things. That sounds obvious until you compare it with decades of advertising built to say things.

SOS POS for BCP turned payment terminals into phone-theft recovery and account-blocking points. Data was used as protection, not just measurement.

Duobell for Škoda solved a real cycling safety problem with a bell designed to cut through noise-cancelling headphones, then shared the blueprint openly.

AXA France’s Three Words, covered in Cannes reporting as the Creative Effectiveness Grand Prix winner, changed policy language around domestic violence. The idea was not trapped inside the campaign. It entered the product promise.

Heineken’s Could Have Been a Heineken turned voice notes into an argument for meeting people in person. The idea used a digital habit to pull people back into physical social life.

These are not simply messages. They are interventions. They protect, alter, help, reconnect, prove, or repair something.

This is a hard lesson for AI marketing. If the model is only used to generate statements, the work stays thin. The better question is what the brand can make easier, safer, faster, more social, more useful, or more memorable.

The winners and standout work point in the same direction:

WorkWhat the jury and industry seemed to rewardSignal
SOS POS for BCPA normal payment terminal became a phone-theft recovery and account-blocking tool.Data as public utility, not reporting.
Build Your Own Super Bowl for Uber EatsThe app became a participatory media surface where fans made their own Super Bowl commercial.Product as campaign platform.
Uva Uva Bombón for UVA AppThe brand moved fast around Bad Bunny and the Super Bowl instead of waiting for a long planning cycle.Cultural timing as creative advantage.
The KitKat HeistA real theft crisis became a public participation moment.Crisis as choreographed communication.
Could Have Been a HeinekenVoice notes became a prompt for in-person connection.Digital habit redirected into human ritual.
Expedition Impossible for Columbia SportswearA niche belief community became a stage for proving product durability.Subculture as media.
Duobell for ŠkodaA safety problem became a practical product idea with an open blueprint.Usefulness plus shareability.
Vaseline OriginalsCommunity product hacks became a way to renew a heritage brand.Creator culture as brand memory.

This is the pattern we have been circling for months. In The Next Chapter in Marketing, we argued that AI shifts marketing from campaign optimisation toward helping people achieve goals. In The Empowerment Imperative, we framed the fork more sharply: the same agentic machinery can extract attention or expand human capability. In Agentic Commerce, the practical rule was simpler: design for trust and verification, not persuasion.

Cannes did not use exactly that language. The winners made the point anyway.

At Cannes 2026, utility was not a dull cousin of creativity. Utility was the creative act.

Discovery 4: Product Became Media

The old campaign model sent people from media to product. The new model often starts inside the product.

Uber Eats’ Build Your Own Super Bowl let fans create personalised Super Bowl commercials through the app. The product became the campaign surface. The app was not just a destination after the media spend. It was the media.

SOS POS used payment terminals as a communications and protection layer.

KitKat’s The KitKat Heist reportedly transformed a real theft crisis into a public participation moment. The crisis became the medium.

Vaseline’s Originals reactivated a heritage product through the community of people who had already been inventing uses for it online. The product’s informal culture became the campaign material.

This matters because AI-native marketing will not stop at asset generation. Agents, apps, product interfaces, loyalty flows, checkout surfaces, packaging, customer service, creator channels, and data signals are becoming campaign surfaces.

Media is no longer only where the campaign runs. Media is what the brand can operate.

That is a bigger change than AI copywriting.

Discovery 5: Data Became A Creative Material

For years, data sat after the idea. The campaign ran. The dashboard reported. The case study proved. The deck turned behaviour into charts.

Cannes 2026 pointed to a different sequence. Data can now sit before and inside the idea.

In SOS POS, data is not a reporting layer. It is the protection mechanism. In product-led and creator-led work, audience behaviour shapes the execution before the media plan is final. In AI-assisted creative systems, feedback loops can alter the next brief before they alter the next optimisation round.

This changes the role of data teams. They are no longer only measurement teams. They become part of creative design.

That does not mean every idea should be data-driven in the deadening sense. It means data can become a material: something shaped, constrained, interpreted, and used to make the experience work.

The danger is obvious. If data only optimises toward what already performs, the brand becomes narrower. If data informs judgment, the brand can learn without collapsing into sameness.

That line is thin. Cannes seemed to reward the teams that knew where it was.

Discovery 6: Speed Became Part Of The Idea

Some of the strongest Cannes examples were not simply big. They were fast.

Uva Uva Bombón for UVA App used a cultural moment around Bad Bunny and the Super Bowl to drive immediate challenger-brand action. The KitKat work turned a real incident into a managed cultural moment. Creator-led marketing runs on similar physics: notice, respond, publish, learn.

For years, brands talked about cultural relevance as if culture waited for approval. It does not.

The old planning cycle was built for controlled campaigns:

Research
↓
Brief
↓
Creative development
↓
Production
↓
Launch
↓
Measurement

The faster world behaves differently:

Signal
↓
Interpretation
↓
Decision
↓
Execution
↓
Feedback
↓
Next signal

AI can help with this. It can monitor, summarise, draft, version, and package. But speed without judgment creates accidents. Speed without brand memory creates opportunism. Speed without approval design creates risk.

The future belongs to teams that can move quickly without becoming careless. That is an operating-system problem.

Discovery 7: Creators Became Strategic Infrastructure

Cannes was also a creator festival. Business Insider reported that more than 250 creators were expected to attend, with creator marketing gaining dedicated tracks, spaces, and executive attention. The language is changing too. Some people now prefer creator commerce because creators increasingly shape products, distribution, feedback, and cultural meaning.

This is not influencer marketing with better lighting. Creators are becoming strategic infrastructure.

They carry audience trust. They understand formats before brands do. They receive feedback every day. They know what people skip, share, mock, imitate, and buy. They move at a speed most organisations cannot match.

That creates tension. Brands want control. Creators need voice. Brands want consistency. Culture rewards specificity. Brands want safety. Attention often begins at the edge.

The strongest teams will treat creators less like rented reach and more like distributed cultural intelligence. That requires different workflows, different approvals, different briefs, and different measures of success.

It also requires humility. A creator is not a media slot. A creator is a person with an audience that has already chosen them.

The Creative Operating System

The pattern underneath Cannes 2026 is clearer once the campaign examples are stripped back. The winning organisations are not just producing better assets. They are building better conditions for judgment.

Call it a creative operating system.

Culture
↓
Brand memory
↓
Human judgment
↓
AI orchestration
↓
Creative evaluation
↓
Product, media, and creator execution
↓
Market feedback
↓
Next campaign

This is where the Anti-Slop Stack meets the Cannes evidence. The stack was a way of asking what brands need when content becomes abundant. Cannes showed those needs in the wild.

Memory appeared in heritage brands using their own communities and histories. Utility appeared in work that solved real problems. Cultural timing appeared in fast opportunistic campaigns. Evaluation appeared in the festival itself. Creators appeared as distributed taste and distribution systems. AI appeared as infrastructure, useful but insufficient.

The creative operating system is not a dashboard. It is the set of habits, tools, standards, memories, approval loops, collaborators, and feedback mechanisms that decide whether production becomes valuable.

This is the real Cannes signal. Creativity is becoming less like a department and more like an organisational capability.

What This Means For AI-Native Marketing

AI-native marketing should not mean generating more ads with fewer people. That is the least interesting version. The Cannes version is harder and better.

An AI-native marketing system should help a team:

  • sense cultural signals in real time
  • retrieve brand memory before generating work
  • evaluate ideas against distinctiveness and long-term identity
  • coordinate product, media, creators, owned channels, and commerce surfaces
  • turn data into creative material before it becomes reporting
  • produce useful experiences rather than only messages
  • keep humans in the loop for taste, risk, and strategic judgment
  • learn from the market without chasing every short-term signal

This is a very different brief from “make me ten posts.” It is closer to building a marketing lab.

The agent does not replace the creative team. It helps the team see more, remember more, test more, and move with less operational drag.

The human role becomes sharper: decide what matters, what fits, what should be killed, and when the brand should move.

What Marketers Should Take From Cannes

Start with the work, then ask uncomfortable questions.

If your brand won tomorrow, would it win for a campaign or for a system? Do you know which ideas your brand should never make? Can your team explain what makes work recognisably yours without quoting the brand book?

Do your AI workflows retrieve brand memory before generating content? Do your review meetings separate fluent work from distinctive work? Do creators have room to shape the idea, or only distribute it?

Can your product become a campaign surface? Can your data become a creative material? Can your organisation respond to culture before the moment is gone?

These are not Cannes questions. They are operating questions. Cannes simply made them visible.

What Engineers Should Build

The engineering brief is becoming clearer: build systems that make judgment easier to apply.

That means:

  • brand-memory retrieval over approved assets, past campaigns, claims, and examples
  • structured briefs with objectives, behavioural hypotheses, constraints, and approval status
  • creative skill libraries for repeatable workflows
  • evaluation harnesses that score brand fit, distinctiveness, behavioural plausibility, and compliance
  • provenance logs for model, prompt, skill version, source material, rights, and approver
  • feedback models that connect creative metadata to market outcomes
  • creator collaboration workflows with review boundaries that preserve voice
  • product and commerce surfaces that can become campaign interfaces
  • human approval gates for high-risk claims, cultural moments, and paid deployment

The best AI marketing systems will look less like content generators and more like creative control planes. They will know the brand, know the context, generate, and then ask whether the work deserves to exist.

Build The Operating System

If you want to turn this argument into something your team can actually use, we have started publishing the practical implementation layer here:

The new Creative Operating System pack includes starter skills for auditing brand memory, designing utility-led campaign concepts, mapping product-as-media opportunities, triaging cultural timing, briefing creators, and evaluating AI-generated work before launch.

It is not a finished doctrine. It is a reference implementation: a way to make the Cannes lesson operational. If AI makes content easier to produce, the advantage moves to the operating system around the content.

The Answer Cannes Gave Us

The biggest surprise of Cannes Lions 2026 was not that AI dominated the festival. It was that AI almost disappeared into the machinery.

The interesting work happened above it: in judgment, utility, cultural timing, product surfaces, creators, and the organisations capable of making these pieces move together.

A week before Cannes, we asked what AI could not replace. Cannes gave a practical answer.

It cannot replace the ability to recognise what matters before everyone else does.

It cannot replace the memory that keeps a brand from dissolving into the category average.

It cannot replace the taste that kills the acceptable idea.

It cannot replace the social intelligence that knows when a product, a creator, a crisis, or a piece of data can become the medium.

AI will be in all of this. Of course it will.

The brands that use it well will not be the ones that generate the most. They will be the ones that build the best machinery for deciding what to keep.

That has always been the game. Cannes just made the rules easier to see.

Reference Reading

Other Articles