By Nicholas Ward, President & Co-Founder, Koddi
Commerce media is entering a new phase. After years of rapid growth, it now accounts for around one in every five dollars spent by U.S. advertisers. Brands have been drawn to commerce media for its targeting, ability to drive incremental sales and robust reporting. These are the playing fields on which the various platforms and networks have competed, until now. As AI becomes deeply embedded throughout the shopper journey, a new factor is becoming critical. And it’s one few people talk about: infrastructure.
Infrastructure is not a subject to get hearts racing. Tech stacks, servers and data retrieval systems can’t match the flair and excitement of the latest AI model. But they are the systems on which AI and automation depend. Even the best sports car can only go as fast as the road conditions allow.
But the demands on that infrastructure have been growing, and they are about to increase exponentially as AI advances. Within a year or two, the invisible technology at work behind the scenes will be crucial in determining which commerce media platforms thrive.
A Need for Speed
Advantage in commerce media is measured in milliseconds. For years, the industry’s rule of thumb has been that every 100 ms delay in serving an ad causes conversion to drop by 7%. But as websites and apps get even faster, ad servers have to keep pace. Ads served slower than the page around them degrade the user experience and get deprioritized. Meeting advertiser and consumer expectations requires low latency infrastructure that can deliver real-time ad decisioning at scale. Building that requires detailed planning and smart execution.
But in a fast-growing industry like commerce media, some platforms prioritized what worked in the moment rather than building a solid foundation. Their teams grabbed the technology they needed and stitched it together, adding a campaign management layer here, a reporting dashboard there and some analytics on the side. That worked for a while, but AI will force a shakeout in the industry, with platforms that are held together by workarounds and fixes quickly coming unstuck.
The reason is the unforgiving math around AI. Traditional systems often make one decision, serve one report or answer one query at a time. Agentic AI workflows issue chains of queries, each informed by the previous step. Even small excess latencies compound into noticeable delays.
A power user running a traditional dashboard might pull a dozen reports a day. With an AI agent, that same user could generate hundreds. On a solid tech stack, that’s an advantage. But on an improvised one, it becomes a headache.
Building for Scale
A robust platform has to be built to scale from the start. That means carefully considered architecture, low-latency ad serving, API integrations and clean data. That work pays off in a network that can adapt, add channels and give partners the flexibility their business needs as it develops. For anyone evaluating a commerce media tech partner, it’s worthwhile to look past the pitch and scrutinize the infrastructure beneath it.
One network we worked with took this approach. They thought about scaling from day one and focused intently on the underlying technology. Operating in a complex industry with millions of products, multiple geographies and highly variable demand, they asked from the outset how to create a seamless experience for advertisers as they grew. This led to the integration of additional automation and AI systems designed to operate at scale. One notable finding from that work: when advertisers received an AI-powered response in seconds rather than waiting for a human callback, their satisfaction scores were higher. Speed is its own form of service. But it only works when the tech beneath it can keep up.
A solid commerce media platform should simplify things with configurable workflows for advertisers of any size, reporting with genuine depth and support for automation and conversational AI. That all requires strong data infrastructure and careful design from the ground up.
The Decision Economy Is Already Here
Commerce media is moving from a world where value lives in visible placements like a banner ad or a search result to one where it is in influence over decisions made by AI agents. That requires faster and more agile decisioning, new measurement techniques, and sophisticated AI orchestration layers.
Advertisers are already moving their budgets from paid search to agentic AI. Some are testing the waters with tens of thousands of dollars in spending, and a few front runners are already committing six- or even seven-figure sums. But the direction of travel is clear. Autonomous, decision-making agents are reshaping commerce media and will soon be deeply embedded in bidding, optimization and campaign execution.
Theorizing about these changes is all very well. But the opportunity belongs to platforms and networks that think carefully about their infrastructure, and build it.
