Agentic AI Won’t Kill Premium Media, but It Will Make It Prove Its Value
CANNES, France — Artificial intelligence may soon handle more media buying, campaign optimization and audience targeting without needing a human to hover over every transaction. It still hasn’t figured out how to make the advertising industry simple, which may be its most difficult assignment.
Executives from Samsung Ads, Seven.One Media, Accenture Song and MNTN discussed how buyers and sellers can protect premium advertising as autonomous AI systems take on more work. Darline Jean, managing director of global publisher platforms at Google, moderated the panel at the Cannes Lions International Festival of Creativity.
The speakers agreed that agentic technology will significantly change advertising. They didn’t entirely agree on whether the industry has already arrived at that future or is merely approaching the driveway while arguing about data permissions.
Mark Douglas, president and chief executive of MNTN, said automation is already deeply embedded in advertising platforms.
“If we’re talking about long tail of advertisers, we’re already there,” Douglas said. “It’s highly automated to do advertising on social, search, on streaming TV if you’re doing direct response marketing.”
For large brands and agencies, he argued, technology isn’t the main obstacle.
“It’s not about the technology, it’s about the people,” Douglas said.
That may come as a relief to the machines, which can now blame humans for delayed adoption just as humans blame technology for almost everything else.
Premium inventory gets a second life
Douglas doesn’t expect automated buying to turn every advertisement into an interchangeable commodity. He predicted that brands will use agentic systems as the foundation of their marketing while becoming even more interested in premium content and distinctive placements.
Brands will seek “premium opportunities, premium content, unique placements” that help them stand apart, he said.
MNTN already helps smaller and midsize advertisers secure placements around programming such as the World Cup, the Final Four and major television premieres. These marketers want measurable returns, but they also understand that placing an advertisement somewhere memorable can do more for a brand than winning an automated auction by three decimal places.
Douglas said advertisers are still asking how premium content can “help differentiate my brand.”
“I think both are going to happen at the same time,” he said.
In other words, automation may standardize the machinery of advertising without standardizing every marketing decision. Machines can optimize the media plan, but they may struggle to explain why an executive suddenly feels a profound strategic connection to an awards-show sponsorship that includes excellent seats.
Samsung sees unfinished plumbing
Youssef Ben-Youssef, head of ads business development, programmatic and privacy at Samsung Ads, offered a more cautious assessment.
Samsung operates premium connected television inventory, including advertisements within programming and native placements on television home screens. Ben-Youssef said the infrastructure needed to buy those opportunities through autonomous agents isn’t fully ready.
“The pipes to be able to execute campaigns and buy that kind of inventory are not agentic-enabled yet,” he said.
Samsung is beginning to test transactions between agents on the buy side and sell side. Those experiments should help the industry learn how to standardize the process and introduce agentic buying more broadly.
The work also raises privacy questions. A publisher may want to share detailed audience signals so buyers understand the value of an impression. However, once an agent receives those signals, the publisher needs assurance that the information won’t be retained or used for another purpose.
“They can use it to buy the impression, but who knows what they can do with it afterwards?” Ben-Youssef said.
That question captures the central dilemma. Publishers want agents to know enough to pay premium prices, but not so much that an automated buyer leaves the transaction carrying the family silver and a detailed profile of everyone in the house.
Context gives automation something to think about
Sascha Stojakovic, senior vice president of technology and partnerships at Seven.One Media, said AI can make premium television content more useful to advertisers by extracting richer contextual signals.
Seven.One uses AI to analyze more than keywords. Its systems can identify emotions, moods and other characteristics within individual scenes. Those insights can be combined with first-party audience information while respecting Europe’s stricter privacy requirements.
Stojakovic said simple factors such as price and reach don’t require sophisticated agents.
“I don’t need agents for that, to be honest. It’s just a simple algorithm,” he said. “But if I enrich it with contextual data, then it becomes a thing.”
That richer information can help buying systems understand why one advertising opportunity may be more valuable than another. It also gives publishers a way to distinguish premium programming from the vast supply of digital content that appears to have been created during someone’s lunch break.
Still, Stojakovic warned against adopting agentic systems merely to claim an industry first.
“Just for the sake of being the first doing that, it’s not what we want,” he said.
Seven.One is experimenting, learning from failures and refining its approach. Stojakovic acknowledged that the technology is advancing so rapidly that today’s confident prediction could look ridiculous within weeks. In advertising, that qualifies as unusual candor and possibly an operational risk.
Agents need meaning, memory and guardrails
Arun Kumar, global head of data and AI strategy and partnerships at Accenture Song, distinguished true agents from conventional automation.
“An agent is where you have some level of autonomy, there’s some level of memory and there’s some level of reasoning,” Kumar said.
The larger opportunity is to connect intelligence across business functions. Customer service, supply chains, marketing and sales often operate with separate information. That’s why a customer can complain about a product one day and receive an advertisement for the same item the next.
“It’s because they have not connected their databases,” Kumar said.
Agentic systems could break down those silos. To do that safely, agents need context about a company’s objectives, customers and brand identity. A marketing agent can’t act with meaningful autonomy if it doesn’t know how the brand should respond in different situations.
“An agent needs to understand what is true brand voice,” Kumar said.
Brands will also need to share some of that context with publishers and other partners. Otherwise, agents on opposite sides of a transaction may communicate efficiently while misunderstanding what either company actually wants, recreating a familiar advertising-industry experience at machine speed.
Humans retain the creative premium
Kumar predicted that agents will be “quite ruthless” about pursuing measurable objectives. They won’t approve a sponsorship because an executive wants to attend the event or because the hospitality suite has a particularly persuasive view.
That discipline could create a new performance baseline across the industry. As brands adopt similar tools and reach similar levels of technical sophistication, competitive advantage will depend on what they do with the resulting intelligence.
Kumar expects humans to retain an important role above that automated foundation.
“There will always be a premium paid for creative ideas that come from humans,” he said.
He also doesn’t expect creative directors to disappear soon because agents can’t yet reproduce the judgment that produces genuinely distinctive ideas.
The agentic future therefore may not eliminate buyers, sellers or creative professionals. It may eliminate some repetitive work while forcing every participant to demonstrate what it contributes beyond operating complicated machinery.
That should be good news for premium media. It may be slightly less comforting for anyone whose main contribution was explaining why the machinery had to remain complicated.
You’re watching The Beet.TV Leadership Sessions at Cannes Lions 2026, presented by Google Ad Manager. For more videos from this series, please visit this page. You can find all of our coverage from Cannes Lions 2026 here.
CANNES, France — Artificial intelligence may soon handle more media buying, campaign optimization and audience targeting without needing a human to hover over every transaction. It still hasn’t figured out how to make the advertising industry simple, which may be its most difficult assignment.
Executives from Samsung Ads, Seven.One Media, Accenture Song and MNTN discussed how buyers and sellers can protect premium advertising as autonomous AI systems take on more work. Darline Jean, managing director of global publisher platforms at Google, moderated the panel at the Cannes Lions International Festival of Creativity.
The speakers agreed that agentic technology will significantly change advertising. They didn’t entirely agree on whether the industry has already arrived at that future or is merely approaching the driveway while arguing about data permissions.
Mark Douglas, president and chief executive of MNTN, said automation is already deeply embedded in advertising platforms.
“If we’re talking about long tail of advertisers, we’re already there,” Douglas said. “It’s highly automated to do advertising on social, search, on streaming TV if you’re doing direct response marketing.”
For large brands and agencies, he argued, technology isn’t the main obstacle.
“It’s not about the technology, it’s about the people,” Douglas said.
That may come as a relief to the machines, which can now blame humans for delayed adoption just as humans blame technology for almost everything else.
Premium inventory gets a second life
Douglas doesn’t expect automated buying to turn every advertisement into an interchangeable commodity. He predicted that brands will use agentic systems as the foundation of their marketing while becoming even more interested in premium content and distinctive placements.
Brands will seek “premium opportunities, premium content, unique placements” that help them stand apart, he said.
MNTN already helps smaller and midsize advertisers secure placements around programming such as the World Cup, the Final Four and major television premieres. These marketers want measurable returns, but they also understand that placing an advertisement somewhere memorable can do more for a brand than winning an automated auction by three decimal places.
Douglas said advertisers are still asking how premium content can “help differentiate my brand.”
“I think both are going to happen at the same time,” he said.
In other words, automation may standardize the machinery of advertising without standardizing every marketing decision. Machines can optimize the media plan, but they may struggle to explain why an executive suddenly feels a profound strategic connection to an awards-show sponsorship that includes excellent seats.
Samsung sees unfinished plumbing
Youssef Ben-Youssef, head of ads business development, programmatic and privacy at Samsung Ads, offered a more cautious assessment.
Samsung operates premium connected television inventory, including advertisements within programming and native placements on television home screens. Ben-Youssef said the infrastructure needed to buy those opportunities through autonomous agents isn’t fully ready.
“The pipes to be able to execute campaigns and buy that kind of inventory are not agentic-enabled yet,” he said.
Samsung is beginning to test transactions between agents on the buy side and sell side. Those experiments should help the industry learn how to standardize the process and introduce agentic buying more broadly.
The work also raises privacy questions. A publisher may want to share detailed audience signals so buyers understand the value of an impression. However, once an agent receives those signals, the publisher needs assurance that the information won’t be retained or used for another purpose.
“They can use it to buy the impression, but who knows what they can do with it afterwards?” Ben-Youssef said.
That question captures the central dilemma. Publishers want agents to know enough to pay premium prices, but not so much that an automated buyer leaves the transaction carrying the family silver and a detailed profile of everyone in the house.
Context gives automation something to think about
Sascha Stojakovic, senior vice president of technology and partnerships at Seven.One Media, said AI can make premium television content more useful to advertisers by extracting richer contextual signals.
Seven.One uses AI to analyze more than keywords. Its systems can identify emotions, moods and other characteristics within individual scenes. Those insights can be combined with first-party audience information while respecting Europe’s stricter privacy requirements.
Stojakovic said simple factors such as price and reach don’t require sophisticated agents.
“I don’t need agents for that, to be honest. It’s just a simple algorithm,” he said. “But if I enrich it with contextual data, then it becomes a thing.”
That richer information can help buying systems understand why one advertising opportunity may be more valuable than another. It also gives publishers a way to distinguish premium programming from the vast supply of digital content that appears to have been created during someone’s lunch break.
Still, Stojakovic warned against adopting agentic systems merely to claim an industry first.
“Just for the sake of being the first doing that, it’s not what we want,” he said.
Seven.One is experimenting, learning from failures and refining its approach. Stojakovic acknowledged that the technology is advancing so rapidly that today’s confident prediction could look ridiculous within weeks. In advertising, that qualifies as unusual candor and possibly an operational risk.
Agents need meaning, memory and guardrails
Arun Kumar, global head of data and AI strategy and partnerships at Accenture Song, distinguished true agents from conventional automation.
“An agent is where you have some level of autonomy, there’s some level of memory and there’s some level of reasoning,” Kumar said.
The larger opportunity is to connect intelligence across business functions. Customer service, supply chains, marketing and sales often operate with separate information. That’s why a customer can complain about a product one day and receive an advertisement for the same item the next.
“It’s because they have not connected their databases,” Kumar said.
Agentic systems could break down those silos. To do that safely, agents need context about a company’s objectives, customers and brand identity. A marketing agent can’t act with meaningful autonomy if it doesn’t know how the brand should respond in different situations.
“An agent needs to understand what is true brand voice,” Kumar said.
Brands will also need to share some of that context with publishers and other partners. Otherwise, agents on opposite sides of a transaction may communicate efficiently while misunderstanding what either company actually wants, recreating a familiar advertising-industry experience at machine speed.
Humans retain the creative premium
Kumar predicted that agents will be “quite ruthless” about pursuing measurable objectives. They won’t approve a sponsorship because an executive wants to attend the event or because the hospitality suite has a particularly persuasive view.
That discipline could create a new performance baseline across the industry. As brands adopt similar tools and reach similar levels of technical sophistication, competitive advantage will depend on what they do with the resulting intelligence.
Kumar expects humans to retain an important role above that automated foundation.
“There will always be a premium paid for creative ideas that come from humans,” he said.
He also doesn’t expect creative directors to disappear soon because agents can’t yet reproduce the judgment that produces genuinely distinctive ideas.
The agentic future therefore may not eliminate buyers, sellers or creative professionals. It may eliminate some repetitive work while forcing every participant to demonstrate what it contributes beyond operating complicated machinery.
That should be good news for premium media. It may be slightly less comforting for anyone whose main contribution was explaining why the machinery had to remain complicated.
You’re watching The Beet.TV Leadership Sessions at Cannes Lions 2026, presented by Google Ad Manager. For more videos from this series, please visit this page. You can find all of our coverage from Cannes Lions 2026 here.
