Google’s Peentoo Patel Sees AI Turning Adtech Hours Into Minutes
CANNES, France – Artificial intelligence may someday buy a Super Bowl commercial without breaking into a digital sweat. For now, Google is giving it a slightly less glamorous assignment: figuring out why an ad campaign stopped delivering before everyone finishes blaming everyone else.
Peentoo Patel, senior director of product management at Google, discussed the company’s development of AI-powered advertising tools during a fireside conversation with Mike Shields, founder of Next in Media, at the Cannes Lions International Festival of Creativity.
Publishers generally are enthusiastic about AI’s potential, Patel said. They’re exploring how the technology can improve routine advertising operations, support new products and change how media is bought and sold.
“The technology has a lot of potential,” Patel said. “How can we harness AI technology, bring new products to market?”
That potential stretches from analyzing content to troubleshooting campaigns and, eventually, coordinating transactions among buyers and sellers. The advertising industry has spent years adding layers of machinery to media buying. Now it’s asking AI to locate the instruction manual.
Ask Ad Manager joins the conversation
Google introduced Ask Ad Manager, which brings its Gemini models into the Google Ad Manager platform. The tool marks a shift from reactive AI features toward what Patel described as a proactive, conversational agent.
Ad-operations teams often start their mornings by reviewing underperforming line items. They then dig through reports, logs, emails and creative files to diagnose the problem. It’s the ad-tech version of looking for a missing sock, except the sock contains somebody’s quarterly revenue target.
Ask Ad Manager is designed to let users ask why a particular line item isn’t delivering. The system examines data in the platform and suggests possible causes.
“So, the idea is, can we go from hours to minutes to making sure you can recover that revenue and spend less time trying to really troubleshoot,” Patel said.
The agent also can assist with more complicated reporting. It might compare one bidder’s performance with others in the system, then recommend changes involving pricing or deal management. Deep links can direct users to the appropriate area of Ad Manager to make those changes.
Google also is developing specialized agents for advertisers, agencies and publishers. These could support inventory discovery, pricing and campaign execution. The company also wants first-party and third-party agents to connect with Ad Manager at scale.
AI looks for room inside the picture
Patel said AI could help publishers create advertising formats that complement their existing inventory without exhausting viewers.
“Imagine a world where you can put maybe an overlay or an in-scene banner on an open scene for video or image,” he said.
A better understanding of video content could help a system identify moments when an overlay or another placement would be relevant and less intrusive. That could create more inventory without simply adding another commercial break and daring viewers to remain conscious.
This opportunity is especially important in connected television and livestreaming. Those environments bring technical complications including device fragmentation, varying streaming standards and the pressure to deliver ads during high-profile events.
“Live is very complicated,” Patel said. “You have a very specific point in time with high-stakes events where you need to make sure your streams deliver, your ads deliver.”
Google’s Dynamic Ad Insertion technology has supported events with millions of simultaneous streams. Patel cited the Cricket World Cup, which reached 11 million concurrent devices. As he and Shields noted, cricket may not be the soccer World Cup, but it’s hardly a neighborhood match with three spectators and a folding chair.
More advertisers bring more creative headaches
AI-generated advertising could help broadcasters attract smaller businesses and performance marketers that lack the resources to produce television commercials. Lowering the cost of creative production could open premium video inventory to a wider group of buyers.
It also could unleash a parade of questionable ads into environments where broadcasters maintain strict standards. Networks don’t want to discover during a championship match that an AI-generated commercial contains a strange claim, a mangled logo or a human hand with the inventory capacity of a baseball team.
Google is using AI-powered creative-classification models to examine ads for brand-safety and compliance issues. Patel said the company analyzed more than 350 million unique creatives during the past year.
The system evaluates landing pages, images, video and JavaScript rather than applying only broad classifications. Automating that review can help publishers expand their demand partnerships without assigning a human being to inspect every creative.
“By reducing false positives, we’ve seen 40% increase in buyer spend in key categories,” Patel said.
Agents aren’t buying the Super Bowl yet
The largest question is whether advertisers eventually will trust autonomous agents to spend money and execute media buys. Patel said the technology remains in an experimental stage, particularly when agents interact with publisher inventory.
Direct deals could offer an early testing ground because publishers can establish guarantees and controls inside their ad servers. Lower-stakes transactions also could let buyers and sellers experiment before handing an agent the keys to a nine-figure media budget.
“There are lower-stakes places you could start with sort of experimenting with this and making sure all the right controls are in place,” Patel said.
That approach leaves room for humans to set restrictions, approve creative and decide which inventory can be sold. The agent may eventually make the purchase, but for now it’ll remain under supervision.
As Shields joked, the logical endpoint arrives when an agent buys a Super Bowl ad and renders the rest of the advertising industry unnecessary. Until then, it can begin by rescuing underperforming line items and sparing ad-operations teams from another morning spent interrogating a spreadsheet.
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 someday buy a Super Bowl commercial without breaking into a digital sweat. For now, Google is giving it a slightly less glamorous assignment: figuring out why an ad campaign stopped delivering before everyone finishes blaming everyone else.
Peentoo Patel, senior director of product management at Google, discussed the company’s development of AI-powered advertising tools during a fireside conversation with Mike Shields, founder of Next in Media, at the Cannes Lions International Festival of Creativity.
Publishers generally are enthusiastic about AI’s potential, Patel said. They’re exploring how the technology can improve routine advertising operations, support new products and change how media is bought and sold.
“The technology has a lot of potential,” Patel said. “How can we harness AI technology, bring new products to market?”
That potential stretches from analyzing content to troubleshooting campaigns and, eventually, coordinating transactions among buyers and sellers. The advertising industry has spent years adding layers of machinery to media buying. Now it’s asking AI to locate the instruction manual.
Ask Ad Manager joins the conversation
Google introduced Ask Ad Manager, which brings its Gemini models into the Google Ad Manager platform. The tool marks a shift from reactive AI features toward what Patel described as a proactive, conversational agent.
Ad-operations teams often start their mornings by reviewing underperforming line items. They then dig through reports, logs, emails and creative files to diagnose the problem. It’s the ad-tech version of looking for a missing sock, except the sock contains somebody’s quarterly revenue target.
Ask Ad Manager is designed to let users ask why a particular line item isn’t delivering. The system examines data in the platform and suggests possible causes.
“So, the idea is, can we go from hours to minutes to making sure you can recover that revenue and spend less time trying to really troubleshoot,” Patel said.
The agent also can assist with more complicated reporting. It might compare one bidder’s performance with others in the system, then recommend changes involving pricing or deal management. Deep links can direct users to the appropriate area of Ad Manager to make those changes.
Google also is developing specialized agents for advertisers, agencies and publishers. These could support inventory discovery, pricing and campaign execution. The company also wants first-party and third-party agents to connect with Ad Manager at scale.
AI looks for room inside the picture
Patel said AI could help publishers create advertising formats that complement their existing inventory without exhausting viewers.
“Imagine a world where you can put maybe an overlay or an in-scene banner on an open scene for video or image,” he said.
A better understanding of video content could help a system identify moments when an overlay or another placement would be relevant and less intrusive. That could create more inventory without simply adding another commercial break and daring viewers to remain conscious.
This opportunity is especially important in connected television and livestreaming. Those environments bring technical complications including device fragmentation, varying streaming standards and the pressure to deliver ads during high-profile events.
“Live is very complicated,” Patel said. “You have a very specific point in time with high-stakes events where you need to make sure your streams deliver, your ads deliver.”
Google’s Dynamic Ad Insertion technology has supported events with millions of simultaneous streams. Patel cited the Cricket World Cup, which reached 11 million concurrent devices. As he and Shields noted, cricket may not be the soccer World Cup, but it’s hardly a neighborhood match with three spectators and a folding chair.
More advertisers bring more creative headaches
AI-generated advertising could help broadcasters attract smaller businesses and performance marketers that lack the resources to produce television commercials. Lowering the cost of creative production could open premium video inventory to a wider group of buyers.
It also could unleash a parade of questionable ads into environments where broadcasters maintain strict standards. Networks don’t want to discover during a championship match that an AI-generated commercial contains a strange claim, a mangled logo or a human hand with the inventory capacity of a baseball team.
Google is using AI-powered creative-classification models to examine ads for brand-safety and compliance issues. Patel said the company analyzed more than 350 million unique creatives during the past year.
The system evaluates landing pages, images, video and JavaScript rather than applying only broad classifications. Automating that review can help publishers expand their demand partnerships without assigning a human being to inspect every creative.
“By reducing false positives, we’ve seen 40% increase in buyer spend in key categories,” Patel said.
Agents aren’t buying the Super Bowl yet
The largest question is whether advertisers eventually will trust autonomous agents to spend money and execute media buys. Patel said the technology remains in an experimental stage, particularly when agents interact with publisher inventory.
Direct deals could offer an early testing ground because publishers can establish guarantees and controls inside their ad servers. Lower-stakes transactions also could let buyers and sellers experiment before handing an agent the keys to a nine-figure media budget.
“There are lower-stakes places you could start with sort of experimenting with this and making sure all the right controls are in place,” Patel said.
That approach leaves room for humans to set restrictions, approve creative and decide which inventory can be sold. The agent may eventually make the purchase, but for now it’ll remain under supervision.
As Shields joked, the logical endpoint arrives when an agent buys a Super Bowl ad and renders the rest of the advertising industry unnecessary. Until then, it can begin by rescuing underperforming line items and sparing ad-operations teams from another morning spent interrogating a spreadsheet.
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.
