Marketers Must Connect Data, AI and Creative to Maximize Results
CANNES, France – Advertisers have spent years collecting data, building clean rooms and producing diagrams that explain how everything theoretically works together. Now comes the awkward part: proving that any of it actually improves advertising.
Denise Colella, vice president of Adobe’s Global Digital Strategy Group, said brands and publishers already have the technology and data needed to personalize advertising across television and other channels. What many still lack is the organizational structure to put those assets to work.
“There’s no shortage of AI systems and there’s no shortage of data, but they live in silos with the brands and with the publishers,” Colella said during a panel at the Cannes Lions International Festival of Creativity.
She joined Ben Hovaness, executive vice president of global specialized services at Omnicom Media, and Peter Randazzo, head of product, ad platforms and operations at NBCUniversal. Mike Shields, founder of Next in Media, moderated the discussion.
Targeting without knowing everyone
The third-party cookie’s prolonged farewell tour has pushed marketers toward first-party data, clean rooms and partnerships that can fill gaps in audience knowledge.
Randazzo said advertisers can’t identify every consumer. They can, however, combine audience data with context and content to infer when someone may be receptive to a message.
“There’s a lot of noise to find that signal,” Randazzo said. “And AI can do a much better job of finding those special moments to reach people.”
That approach moves targeting away from the industry’s long-running fixation on knowing exactly who someone is. Hovaness pointed to Google’s Performance Max and similar offerings from Meta as evidence that AI can find likely customers without starting with a predefined audience.
“I don’t care who you are, as long as I can predict that you’re likely to take a particular action if I serve you a particular ad,” Hovaness said.
It’s an emotionally difficult concept for an industry that has spent untold billions promising advertisers that it knows precisely who everyone is, what they want and possibly what they ate for breakfast. Apparently, a strong prediction may be more valuable than knowing a consumer’s middle name.
Television chases the machines
The panelists also considered whether television could adopt the automated targeting capabilities that have helped Google and Meta capture advertising budgets.
Colella said television companies don’t have much choice.
“Absolutely they should and must try this because, aside from sports, we’re seeing such a degradation in the number of users for TV,” she said.
To preserve audiences and upfront pricing, television providers will need better technology, broader data collaboration and more creative variations. Fortunately, AI can now generate many of those variations without requiring an army of copywriters to debate whether a button should say “Buy now” or “Learn more.”
Colella argued that television still offers something digital platforms struggle to match: the ability to tell emotionally resonant stories at scale.
“Where else do you have that ability to tell the beautiful stories that you can tell with television programming?” she said.
AI can help connect those stories to commerce, allowing television to play a role throughout the funnel instead of waving goodbye to viewers once the credits roll.
Randazzo said NBCUniversal wants to reproduce the automation and accessibility offered by large technology platforms. However, television companies also need advertisers and agencies to share outcomes data so media exposure can be connected with consumer actions.
“We definitely want to replicate the automatic nature of it,” he said. “We definitely want to replicate the ease of it, make it more accessible.”
Creative and media reunite after a long separation
AI also may help reverse the advertising industry’s decades-long separation of creative development from media execution.
“Marketing engines have to be tied to creative engines these days,” Colella said.
She described a connected content supply chain extending from creative ideation through targeting, delivery, analysis and optimization.
“There’s no excuse in today’s world why you wouldn’t connect those systems,” Colella said. “We have the capabilities. We have the AI to do it.”
The industry has dreamed about real-time dynamic creative since at least the late 2000s. Colella recalled working at Right Media when advertising technology companies were already promising personalized creative at scale.
A mere couple of decades, several acquisitions and approximately 47,000 PowerPoint slides later, the pieces may finally be in place.
“We always dreamed about the day when you would be able to do dynamic creative in real time,” Colella said. “We’re here now.”
Advertisers would like to see receipts
Brands are increasingly less impressed by diagrams of agency technology stacks and more interested in whether those systems produce measurable results.
Hovaness said agencies often make similar claims about owning the most data, the best data or the finest marketing platform assembled by humankind. To advertisers, those pitches can start blending together.
“Wonderful schematic. Thank you for showing us a diagram of how your infrastructure’s set up,” he said, summarizing the client reaction. “But anyway, does it work?”
Advertisers want case studies, projections and commitments tied to improvements over existing performance. Some are also becoming more willing to share baseline results, giving agencies and media companies a better chance of demonstrating incremental value.
Randazzo said that change is especially important for television, which traditionally hasn’t had direct access to advertiser outcomes.
“If you’re trying to measure the whole funnel, it’s become more commonplace, more common knowledge that outcomes have to be a part of that conversation,” he said.
The pilot program may now taxi to the gate
For Colella, the biggest obstacles aren’t the availability of data or AI. They’re scattered pilot programs, fragmented contracts and unclear accountability.
“I look at it and I think it’s pilots and governance,” she said.
Companies often have AI projects operating independently across brands, publishers and departments. Colella called for centralized leadership, clear key performance indicators and baselines that let technology providers show whether their systems delivered.
“Let’s get past the pilots,” she said. “Let’s make sure we have leadership that is focused in on AI.”
Hovaness agreed that earlier data collaboration projects were often difficult to implement. Clean rooms generated enormous enthusiasm around 2020 and 2021, but the practitioner experience didn’t always resemble the conference presentation.
“Then you go talk to a practitioner, they’re like, ‘Yeah, we did one. It was terrible. Our clients never want to touch data clean rooms again,’” he said.
AI may lower the technical barriers, especially for multi-party clean rooms. Commercial tensions won’t disappear, of course. Competitors still must decide what information they’re willing to share, with whom and under what conditions.
The technology is getting smarter. Whether the organizations using it can become equally coordinated remains the more complicated experiment.
You’re watching The Beet.TV Leadership Sessions at Cannes Lions 2026, presented by Adobe. For more videos from this series, please visit this page. You can find all of our coverage from Cannes Lions 2026 here.
CANNES, France – Advertisers have spent years collecting data, building clean rooms and producing diagrams that explain how everything theoretically works together. Now comes the awkward part: proving that any of it actually improves advertising.
Denise Colella, vice president of Adobe’s Global Digital Strategy Group, said brands and publishers already have the technology and data needed to personalize advertising across television and other channels. What many still lack is the organizational structure to put those assets to work.
“There’s no shortage of AI systems and there’s no shortage of data, but they live in silos with the brands and with the publishers,” Colella said during a panel at the Cannes Lions International Festival of Creativity.
She joined Ben Hovaness, executive vice president of global specialized services at Omnicom Media, and Peter Randazzo, head of product, ad platforms and operations at NBCUniversal. Mike Shields, founder of Next in Media, moderated the discussion.
Targeting without knowing everyone
The third-party cookie’s prolonged farewell tour has pushed marketers toward first-party data, clean rooms and partnerships that can fill gaps in audience knowledge.
Randazzo said advertisers can’t identify every consumer. They can, however, combine audience data with context and content to infer when someone may be receptive to a message.
“There’s a lot of noise to find that signal,” Randazzo said. “And AI can do a much better job of finding those special moments to reach people.”
That approach moves targeting away from the industry’s long-running fixation on knowing exactly who someone is. Hovaness pointed to Google’s Performance Max and similar offerings from Meta as evidence that AI can find likely customers without starting with a predefined audience.
“I don’t care who you are, as long as I can predict that you’re likely to take a particular action if I serve you a particular ad,” Hovaness said.
It’s an emotionally difficult concept for an industry that has spent untold billions promising advertisers that it knows precisely who everyone is, what they want and possibly what they ate for breakfast. Apparently, a strong prediction may be more valuable than knowing a consumer’s middle name.
Television chases the machines
The panelists also considered whether television could adopt the automated targeting capabilities that have helped Google and Meta capture advertising budgets.
Colella said television companies don’t have much choice.
“Absolutely they should and must try this because, aside from sports, we’re seeing such a degradation in the number of users for TV,” she said.
To preserve audiences and upfront pricing, television providers will need better technology, broader data collaboration and more creative variations. Fortunately, AI can now generate many of those variations without requiring an army of copywriters to debate whether a button should say “Buy now” or “Learn more.”
Colella argued that television still offers something digital platforms struggle to match: the ability to tell emotionally resonant stories at scale.
“Where else do you have that ability to tell the beautiful stories that you can tell with television programming?” she said.
AI can help connect those stories to commerce, allowing television to play a role throughout the funnel instead of waving goodbye to viewers once the credits roll.
Randazzo said NBCUniversal wants to reproduce the automation and accessibility offered by large technology platforms. However, television companies also need advertisers and agencies to share outcomes data so media exposure can be connected with consumer actions.
“We definitely want to replicate the automatic nature of it,” he said. “We definitely want to replicate the ease of it, make it more accessible.”
Creative and media reunite after a long separation
AI also may help reverse the advertising industry’s decades-long separation of creative development from media execution.
“Marketing engines have to be tied to creative engines these days,” Colella said.
She described a connected content supply chain extending from creative ideation through targeting, delivery, analysis and optimization.
“There’s no excuse in today’s world why you wouldn’t connect those systems,” Colella said. “We have the capabilities. We have the AI to do it.”
The industry has dreamed about real-time dynamic creative since at least the late 2000s. Colella recalled working at Right Media when advertising technology companies were already promising personalized creative at scale.
A mere couple of decades, several acquisitions and approximately 47,000 PowerPoint slides later, the pieces may finally be in place.
“We always dreamed about the day when you would be able to do dynamic creative in real time,” Colella said. “We’re here now.”
Advertisers would like to see receipts
Brands are increasingly less impressed by diagrams of agency technology stacks and more interested in whether those systems produce measurable results.
Hovaness said agencies often make similar claims about owning the most data, the best data or the finest marketing platform assembled by humankind. To advertisers, those pitches can start blending together.
“Wonderful schematic. Thank you for showing us a diagram of how your infrastructure’s set up,” he said, summarizing the client reaction. “But anyway, does it work?”
Advertisers want case studies, projections and commitments tied to improvements over existing performance. Some are also becoming more willing to share baseline results, giving agencies and media companies a better chance of demonstrating incremental value.
Randazzo said that change is especially important for television, which traditionally hasn’t had direct access to advertiser outcomes.
“If you’re trying to measure the whole funnel, it’s become more commonplace, more common knowledge that outcomes have to be a part of that conversation,” he said.
The pilot program may now taxi to the gate
For Colella, the biggest obstacles aren’t the availability of data or AI. They’re scattered pilot programs, fragmented contracts and unclear accountability.
“I look at it and I think it’s pilots and governance,” she said.
Companies often have AI projects operating independently across brands, publishers and departments. Colella called for centralized leadership, clear key performance indicators and baselines that let technology providers show whether their systems delivered.
“Let’s get past the pilots,” she said. “Let’s make sure we have leadership that is focused in on AI.”
Hovaness agreed that earlier data collaboration projects were often difficult to implement. Clean rooms generated enormous enthusiasm around 2020 and 2021, but the practitioner experience didn’t always resemble the conference presentation.
“Then you go talk to a practitioner, they’re like, ‘Yeah, we did one. It was terrible. Our clients never want to touch data clean rooms again,’” he said.
AI may lower the technical barriers, especially for multi-party clean rooms. Commercial tensions won’t disappear, of course. Competitors still must decide what information they’re willing to share, with whom and under what conditions.
The technology is getting smarter. Whether the organizations using it can become equally coordinated remains the more complicated experiment.
You’re watching The Beet.TV Leadership Sessions at Cannes Lions 2026, presented by Adobe. For more videos from this series, please visit this page. You can find all of our coverage from Cannes Lions 2026 here.
