Agencies are still selling people hours. That's not going to hold up.
At Cannes Lions 2026, the advertising industry's technology leaders confronted a disorientation that has become its own kind of clarity: artificial intelligence is not arriving — it has already arrived, and the institutions built to sell human creativity are struggling to keep pace with what they have helped unleash. The central tension is not between humans and machines, but between the speed at which opportunity can now be identified and the organizational inertia that prevents anyone from acting on it. What is being negotiated, quietly and urgently, is not just a new set of tools but a new theory of value — what agencies are for, what they sell, and whether the unit of 'people hours' can survive contact with a technology that never sleeps.
- Agency leaders at Cannes are no longer debating whether AI will reshape their industry — they are managing the disorientation of a transformation already underway, leading organizations through change without a reliable map.
- A new professional archetype — the 'orchestrator,' fluent across creative, media, data, and technology — is emerging as the industry's most urgent talent need, yet finding and training these people is proving harder than building the tools they would use.
- Google has displaced OpenAI as the dominant AI provider among agencies, with adoption spreading rapidly into document summarization and research, but concerns over accuracy, bias, legal exposure, and data privacy are creating a ceiling on deeper integration.
- The sharpest bottleneck is not technical but bureaucratic: clients arrive at the moment of action without internal approvals, allocated budgets, or ready assets, leaving agencies unable to deploy the insights they can now surface in hours.
- A proposed structural fix is gaining traction — approving systems rather than individual campaigns, setting the creative and financial playing field in advance so that agility becomes possible within pre-sanctioned boundaries.
- With 61% of agencies still treating AI as a cost of doing business rather than a revenue source, the industry's economic model remains unresolved, and the gap between technological capability and commercial reinvention is widening by the quarter.
The technology leaders who gathered at Cannes Lions this year were not speaking about artificial intelligence as a horizon event. They were speaking about it as the present condition of their industry — and about the particular vertigo that comes from leading organizations through transformation when the answers haven't arrived yet. One executive framed it with unusual candor: how do you project confidence when the learning curves are steep, the existential moments arrive without warning, and your entire organization is watching to see if you know what you're doing?
The conversation kept circling a productive contradiction. Agencies are being asked to move faster, experiment freely, and iterate without fear — while simultaneously being asked to produce work with emotional resonance, craft, and the kind of human meaning that made advertising matter in the first place. The consumer journey is now partly mediated by bots that never sleep, and agility has become the new competitive ground. Both things are true at once, and the industry is still learning to hold them together.
Out of this tension, a new professional role is taking shape. The industry is calling them orchestrators — people who sit at the intersection of creative and media, who understand data, technology, and connected intelligence well enough to make decisions across disciplines without losing themselves in any single one. The concept is clear enough. Finding, training, and building a career path around such people is proving far more difficult than the technology itself.
The deeper bottleneck, however, is organizational rather than human. Agencies can now identify opportunities and surface insights at a speed that would have been unimaginable a few years ago. But when they bring those findings to clients, the machinery stalls. Internal approvals haven't been secured. Budgets haven't been allocated. Assets aren't ready. Governance processes designed for a slower world become the ceiling on what the technology can actually deliver. The appetite for risk, several executives noted, has never been stronger — but appetite and action are separated by institutional friction that no algorithm can dissolve.
One proposed path forward involves a fundamental shift in how approval itself is structured. Rather than approving individual campaigns or specific executions, clients could approve the system — the range of creative possibilities, the floor and ceiling of platform spend, the rules governing how money moves between channels. Approve the playing field, and agility becomes possible within it. The technology can then operate at the speed it was built for.
The numbers reflect an industry in genuine transition. Google has become the leading AI provider among agencies, displacing OpenAI from the prior year. Adoption is broad but uneven — widespread for summarization and research, slower where accuracy, bias, legal risk, and privacy concerns raise the stakes. For AI agents specifically, gaps in expertise and data infrastructure remain significant. And the business model itself is still being written: the majority of agencies still account for AI as a cost rather than a revenue source, and the economic unit of 'people hours' — the foundation on which the industry was built — is increasingly difficult to defend against a technology that does not bill by the hour.
The tech leaders gathered at Cannes Lions this year weren't talking about artificial intelligence as some distant future. They were talking about it as the thing reshaping their business right now—and the disorientation that comes with that speed.
One of them posed a question that cut through the jargon: How do you lead people through this when they're looking to you for answers you don't have yet? The learning curves are steep. Existential moments arrive without warning. You have to figure it out, and fast, and your organization is watching to see if you know what you're doing.
The conversation kept returning to a paradox. Agencies are being asked to move faster than ever, to experiment without fear of failure, to break things and iterate. At the same time, they're being asked to build emotional resonance, to create work that moves people, to maintain the craft that made advertising matter in the first place. One executive put it plainly: we're advertising to bots now, but we're also advertising to humans. Bots never sleep. Agility has become the new competitive ground.
This has created a new kind of job. The industry is calling them orchestrators—people who sit at the intersection of creative and media, who understand data and technology and connected intelligence, who can make decisions across disciplines without getting lost in any single one. They need deep expertise in multiple areas. They need to be masters of their craft while also moving at the speed AI demands. Finding those people, training them, convincing the industry that this is a real career path—that's proving harder than the technology itself.
But the real bottleneck isn't technical. It's organizational. Agencies can identify opportunities faster than ever. They can run experiments at scale. They can surface insights that would have taken weeks to uncover. Then they present those findings to their clients, and the clients can't act. They don't have the internal approvals. They don't have the budget allocated. They don't have the assets ready. They're caught in governance processes designed for a slower world. One executive described the frustration: the appetite for risk-taking on the client side has never been stronger, but when you get to the moment of action, the machinery doesn't move. The client doesn't benefit if the agency can't help them untangle themselves.
The path forward, some argued, is to shift how approval works altogether. Instead of approving individual campaigns or specific creative executions, clients could approve systems—the playing field itself. What's the range of creative possibilities? What are the minimum and maximum spend levels across platforms? How much flexibility do we have to move money between channels? If you approve the system, you can be agile within it. You can move at the speed the technology allows.
The numbers tell a story of a business in transition. Google has become the dominant AI provider among agencies, displacing OpenAI from the previous year. Seventy-four percent of agencies use generative AI to summarize documents and communications. Seventy percent use it for research and competitive intelligence. But accuracy concerns, bias, legal risk, and privacy worries are slowing deeper adoption. For AI agents specifically, the gaps are wider: fifty-four percent of agencies cite lack of expertise, fifty-one percent point to data infrastructure problems.
Sixty-one percent of agencies still treat AI as a cost of business. Thirty-one percent plan to monetize it within the next year. The business model is still being written. One executive from outside the agency world offered a diagnosis that landed hard: agencies are still selling people hours. That's the economic unit they get paid on. And that's not going to hold up over time. The technology is moving faster than the business model can follow.
Citações Notáveis
The challenge that agencies have today is principally that what they're selling today is people hours—that's still the economic unit they get paid on. And that's not going to hold up over time.— Tim Castree, CMO of DoorDash
Agility is the new battleground. We are advertising to bots, but we're also advertising to humans. We still need to have that creativity and innovation.— Agency executive at Cannes Lions
A Conversa do Hearth Outra perspectiva sobre a história
So when they talk about orchestrators, are they describing something that already exists, or are they inventing a new role?
They're inventing it. The closest thing that existed before was maybe a strategic planner or a media director, but this is different. An orchestrator needs to understand creative deeply, media deeply, data, technology, and be able to make decisions across all of it in real time. It's not a role that was necessary when things moved slower.
And they can't find these people?
Not easily. The industry has spent decades training specialists—creative people who know creative, media people who know media. Now you need someone who's a master of multiple crafts. That's a different kind of training, a different kind of career path. They're still figuring out how to build it.
You mentioned clients can't act on opportunities even when agencies find them. What's actually stopping them?
It's not money, usually. It's governance. They need internal approvals for AI-driven work. They don't have the assets ready. They're moving at the speed of their own organization, which is still built for quarterly planning cycles. The agency can move in days. The client needs weeks.
So the agencies are trying to help them move faster?
Yes, but it's tricky. They're suggesting clients approve systems instead of individual plans. Approve the creative range, the budget range, the flexibility to move money around. Then you can be agile within that approved system. It's a different way of thinking about control.
And the technology itself—is that the hard part?
No. That's what's interesting. Everyone says the hard part isn't technical. It's organizational, cultural, structural. The technology works. The business model and the approval processes haven't caught up.