Agentic AI will Put Media Budgets In Motion: Dentsu X’s Brad Stockton
AMENIA, NY — Media buyers have spent years teaching machines to follow their instructions. Agentic artificial intelligence promises to let the machines make more decisions themselves, which sounds wonderfully efficient until someone remembers that the machines will be spending actual money.
Brad Stockton, director of investment at Dentsu X, discussed the emerging technology with Beet.TV contributor David Kaplan at the Beet Retreat Berkshires. Stockton said agentic AI could improve the speed of campaign optimization while helping advertisers allocate impressions across channels and media partners.
“It’s really about the speed in which we’re able to make optimizations into our practice,” Stockton said.
More signals enter the buying equation
Agentic systems are designed to act with greater autonomy than traditional programmatic tools. Instead of waiting for a human to initiate every adjustment, the technology can evaluate signals and respond while a campaign is running.
The potential results will become clearer over the next three, six, nine and 12 months as the technology reaches greater scale, Stockton said. Buyers will examine cross-platform and cross-channel performance to determine the right mix of impressions for each partner.
Agentic AI also could give buyers more data points for deciding where media dollars should flow.
“It’s going to have more data points and signals at our fingertips to then start putting that right impression to the right person at the right time to deliver the right outcome,” Stockton said.
Asked which platforms and formats stand to benefit, Stockton offered the most honest answer available when discussing a rapidly developing technology: “We’re going to find out.”
Partners that deliver the most efficient results should receive more impressions. The machines may be new, but the desire to get more for the client’s money remains comfortably old-fashioned.
Humans keep their hands near the controls
Greater autonomy introduces an obvious question for advertisers: What happens when an AI system makes a poor decision at machine speed?
Stockton said agentic buying will continue to require human intervention as agencies build confidence in the technology. Brand-safety rules must define where advertising can and cannot appear, while exclusion lists and other protections need to be applied correctly.
Humans will audit campaigns before and after they run. That should provide some comfort to brands that do not want an autonomous buying system discovering an exciting new audience in the comments beneath a conspiracy video.
“Today it’s going to be very much manual oversight, with the future really being more about auditing on a consistent basis,” Stockton said.
The long-term vision gives agentic systems more freedom to make and execute buying decisions. For now, however, the robots will have supervisors.
Budgets become more fluid
Agentic buying may require advertisers to loosen rigid channel budgets. Rather than instructing an agency to spend fixed amounts in connected television, social media or display, brands could define the desired outcome and let spending move toward the best-performing channels.
Those outcomes might include awareness, consideration, a specific action or a completed sale. Stockton said agencies must work with clients to establish clear performance indicators and determine the role of each channel within an omnichannel campaign.
“We should be optimizing towards that biggest outcome,” he said.
Media partners that deliver stronger results would be allowed to clear more impressions. That approach could make budgets more responsive, though perhaps less reassuring to anyone emotionally attached to a beautifully color-coded allocation chart.
Getting comfortable with discomfort
The biggest change may be psychological rather than technical. Brands and agencies will have to trust systems that can shift investment across channels instead of following a fixed spending plan.
“The first is we have to get comfortable being uncomfortable,” Stockton said.
That means allowing greater fluidity across omnichannel campaigns while maintaining a clear definition of success. Data and technology can then guide spending toward the most efficient results for clients.
Agentic AI may eventually make media investment faster and more flexible. It will not eliminate the need for judgment. It may simply move humans from pressing every button to checking afterward whether the machine pressed the right ones.
You’re watching coverage from Beet Retreat Berkshires 2026. For more videos from this event, please visit this page.
AMENIA, NY — Media buyers have spent years teaching machines to follow their instructions. Agentic artificial intelligence promises to let the machines make more decisions themselves, which sounds wonderfully efficient until someone remembers that the machines will be spending actual money.
Brad Stockton, director of investment at Dentsu X, discussed the emerging technology with Beet.TV contributor David Kaplan at the Beet Retreat Berkshires. Stockton said agentic AI could improve the speed of campaign optimization while helping advertisers allocate impressions across channels and media partners.
“It’s really about the speed in which we’re able to make optimizations into our practice,” Stockton said.
More signals enter the buying equation
Agentic systems are designed to act with greater autonomy than traditional programmatic tools. Instead of waiting for a human to initiate every adjustment, the technology can evaluate signals and respond while a campaign is running.
The potential results will become clearer over the next three, six, nine and 12 months as the technology reaches greater scale, Stockton said. Buyers will examine cross-platform and cross-channel performance to determine the right mix of impressions for each partner.
Agentic AI also could give buyers more data points for deciding where media dollars should flow.
“It’s going to have more data points and signals at our fingertips to then start putting that right impression to the right person at the right time to deliver the right outcome,” Stockton said.
Asked which platforms and formats stand to benefit, Stockton offered the most honest answer available when discussing a rapidly developing technology: “We’re going to find out.”
Partners that deliver the most efficient results should receive more impressions. The machines may be new, but the desire to get more for the client’s money remains comfortably old-fashioned.
Humans keep their hands near the controls
Greater autonomy introduces an obvious question for advertisers: What happens when an AI system makes a poor decision at machine speed?
Stockton said agentic buying will continue to require human intervention as agencies build confidence in the technology. Brand-safety rules must define where advertising can and cannot appear, while exclusion lists and other protections need to be applied correctly.
Humans will audit campaigns before and after they run. That should provide some comfort to brands that do not want an autonomous buying system discovering an exciting new audience in the comments beneath a conspiracy video.
“Today it’s going to be very much manual oversight, with the future really being more about auditing on a consistent basis,” Stockton said.
The long-term vision gives agentic systems more freedom to make and execute buying decisions. For now, however, the robots will have supervisors.
Budgets become more fluid
Agentic buying may require advertisers to loosen rigid channel budgets. Rather than instructing an agency to spend fixed amounts in connected television, social media or display, brands could define the desired outcome and let spending move toward the best-performing channels.
Those outcomes might include awareness, consideration, a specific action or a completed sale. Stockton said agencies must work with clients to establish clear performance indicators and determine the role of each channel within an omnichannel campaign.
“We should be optimizing towards that biggest outcome,” he said.
Media partners that deliver stronger results would be allowed to clear more impressions. That approach could make budgets more responsive, though perhaps less reassuring to anyone emotionally attached to a beautifully color-coded allocation chart.
Getting comfortable with discomfort
The biggest change may be psychological rather than technical. Brands and agencies will have to trust systems that can shift investment across channels instead of following a fixed spending plan.
“The first is we have to get comfortable being uncomfortable,” Stockton said.
That means allowing greater fluidity across omnichannel campaigns while maintaining a clear definition of success. Data and technology can then guide spending toward the most efficient results for clients.
Agentic AI may eventually make media investment faster and more flexible. It will not eliminate the need for judgment. It may simply move humans from pressing every button to checking afterward whether the machine pressed the right ones.
You’re watching coverage from Beet Retreat Berkshires 2026. For more videos from this event, please visit this page.
Series Videos
