As advertisers push deeper into AI-powered media buying, the industry’s biggest challenge may not be the technology itself, but the quality of the programmatic supply chain underneath it, according to Tyler Romasco, OpenX’s EVP of commercial.

At EMARKETER’s Ad Buyer Strategies Summit, in a conversation with our analyst Yory Wurmser, Romasco argued that marketers are increasingly shifting decision-making closer to publishers in search of better performance, cleaner supply paths, and higher-quality data.

“The fun thing about programmatic is we feel like we’re the original AI when it comes to advertising," Romasco said.

Yory Wurmser and Tyler Romasco speaking at EMARKETER's Future of Digital Summit

Supply-chain inefficiencies remain a major problem

One of the largest issues in programmatic advertising is the complexity and duplication embedded within the supply chain.

“We see in the programmatic ecosystem, up to 72 resellers might touch a piece of inventory before it gets to a marketer,” Romasco said, detailing that OpenX also sees “up to 30x duplication of supply,” which hurts campaign performance and outcomes.

Romasco attributed much of the problem to years of incentives that prioritized scale and low CPMs over transparency and efficiency.

As a result, he said advertisers are becoming more focused on supply quality and direct publisher relationships rather than simply maximizing reach.

Better data and proximity

Romasco argued that moving intelligence closer to publishers gives marketers access to stronger signals and broader visibility into available inventory.

“One of those things that people are getting when they move into the supply side…is you immediately get access to the entire bitstream, all available ad impressions,” he said.

He also emphasized the growing importance of identity resolution and signal quality as advertisers rely more heavily on AI-driven optimization.

“Identity is very fragmented,” Romasco said, adding that stronger identity graphs and higher-quality signals help improve targeting and campaign measurement.

The shift is also increasing pressure on ad tech companies to modernize their infrastructure. Romasco said companies still relying on on-premise systems could struggle to compete as AI workloads expand.

“You need to be 100% a cloud-based company in order to keep up with this accelerated media space,” he said.

The wait for agentic AI

While AI and agentic media buying have become major talking points across the industry, Romasco said fully autonomous agent-to-agent transactions are still several years away.

• Recently, Olyzon and Swivel partnered to develop agent-to-agent buying, with French pharmaceutical company Pierre Fabre Group testing the technology for a US campaign, EMARKETER reported.

For now, most AI applications are focused on improving workflow efficiency and reducing operational friction rather than replacing strategic decision-making, according to Romasco, who cautioned that AI systems are only as effective as the data powering them.

“Garbage in is garbage out,” he said. “If the signal is wrong, if you get the identity wrong, if you get the supply quality wrong, that agent is just going to build and accelerate the bad outcomes.”

Despite the rapid adoption of AI tools, Romasco said marketers still prioritize relationships and strategic expertise alongside automation.

“Agentic is great. AI is great. I can see what it can do, but I buy relationships, and I buy people, and I buy strategy," he said. “We’re not reducing headcount. We’re changing what people do, and we’re refocusing it on the higher value things.”

Watch the full session

We prepared this article with the assistance of generative AI tools and stand behind its accuracy, quality, and originality.