4 steps for differentiating your commerce media network in an AI-first world

For years, commerce media networks have operated under a relatively simple premise: create more opportunities for advertisers to be seen. This happened through offering new sponsored inventory, driving more impressions offsite and in-store, and maximizing clicks through sophisticated relevancy algorithms.

But what happens when consumers stop browsing pages and start asking AI for recommendations? The answer is that visibility becomes less important than influence.

Koddi’s latest research, The state of agentic commerce (media), reveals a fundamental shift underway. As AI increasingly participates in product discovery, research, recommendations, and purchasing decisions, commerce media is moving from a business model based on placements and impressions to one based on access, influence, and outcomes.

This creates a massive change for commerce media networks, forcing them to shift their monetization strategy from attention to decisioning. But it also creates a new opportunity – they can differentiate not by inventory or traffic, but by offering systems that meaningfully impact consumer decisions. Here’s how.

#1: Ditch the impression-based model

Commerce media was built for a world where consumers actively navigated websites, marketplaces, and apps. The value exchange was straightforward: commerce companies had a new revenue stream, advertisers paid for visibility, and consumers got a more relevant and seamless shopping experience. 

But with the introduction of the AI agent, the value shifts entirely. Instead of manually comparing dozens of products, consumers increasingly ask AI systems to do the work for them. The agent researches options, compares trade-offs, narrows the field, and presents recommendations. 

Consumers’ growing dependence on AI agents changes how commerce media networks are able to create value for advertisers and consumers alike. Instead of offering impressions at scale, commerce media networks  should focus on monetizing the AI recommendation and making their products available in a structured way to major LLM platforms.

#2: Build a platform based on access, influence, and execution

Thus, commerce media networks can no longer rely on impressions as their primary source of value. Instead, they must turn their attention to access, influence, and execution.

In the traditional commerce media model, value was created by connecting advertisers with consumers at key moments of discovery. Networks monetized inventory, generated impressions, and optimized placements to drive clicks and conversions.

In an agentic commerce environment, those same outcomes still matter. But the path to achieving them changes dramatically. Instead of competing for placement on a page, brands must compete to be included in an AI agent’s consideration set. Instead of optimizing visibility, networks must optimize how products are discovered, recommended, and purchased through AI-driven experiences.

That starts with access. AI systems need structured, real-time information about products, pricing, availability, promotions, fulfillment options, and consumer preferences. Commerce media networks are uniquely positioned to provide this information because they sit at the intersection of advertisers, product catalogs, and transaction data.

The networks that make their data and inventory accessible to AI systems will be far better positioned than those that treat their commerce experiences as closed ecosystems.

But access alone isn’t enough. Once products are available to an AI agent, the next challenge becomes influence.

AI systems don’t simply display every possible option. They evaluate alternatives, rank products, and generate recommendations. This creates an entirely new opportunity for commerce media networks to help shape consumer decisions.

Just as sponsored placements emerged to influence visibility in search results, new monetization models will surface around recommendation environments. Networks will need mechanisms that allow advertisers to participate in the recommendation process while maintaining relevance and consumer trust.

The final component is execution. Recommendations only create value if they lead to transactions. Consumers must be able to seamlessly move from recommendation to purchase, whether the decision originates on a retailer site, within a commerce app, or through an external AI platform.

Commerce media networks already possess many of the capabilities required to make this possible, including transaction infrastructure, fulfillment systems, measurement capabilities, and advertiser relationships.

These execution capabilities become increasingly important as agentic commerce evolves. Networks that can connect recommendations directly to purchasing outcomes will create significantly more value than those that simply facilitate discovery.

Together, access, influence, and execution represent a new foundation for commerce media.

The networks that thrive won’t be the ones with the most ad placements. They’ll be the ones that make products accessible to AI systems, influence recommendations in meaningful ways, and enable transactions wherever consumers choose to shop.

#3: Get the recommendation engine right

As  impressions become less valuable and recommendations become the primary mechanism for discovery, recommendation quality emerges as one of the most important capabilities a commerce media network can build.

Agentic commerce changes that dynamic as  AI systems increasingly act as intermediaries between consumers and products. Rather than presenting hundreds of options, they narrow the field and recommend a handful of products that best meet a consumer’s needs. In many cases, consumers may never see the full assortment available.

This places enormous importance on the systems that generate recommendations.

Commerce media networks already possess many of the ingredients required to power these experiences. They have access to product catalogs, transaction data, shopper behavior signals, inventory availability, fulfillment information, ratings, reviews, and contextual data about consumer intent. The challenge is bringing these signals together in a way that produces recommendations that consumers trust and advertisers value.

Poor recommendations create risk for everyone involved. Consumers lose confidence in the experience. Advertisers question performance. Networks lose credibility as a trusted intermediary.

Strong recommendations create the opposite effect. Consumers receive better outcomes. Advertisers gain access to highly qualified demand. Networks become increasingly valuable because they influence the products that enter consideration.

The networks that succeed will treat recommendation systems as a core product, not simply a feature. Their ability to connect consumer intent with the right products will increasingly determine their value in the ecosystem.

#4: Optimize for both commerce AI agents and external shopping agents

One of the biggest questions facing the industry is how these decisions will ultimately transpire. 

Will consumers rely primarily on retailer-owned AI experiences? Or, will external AI platforms become the primary interface for shopping?

The answer is likely both. Consumers already trust retailers for product discovery, purchasing, fulfillment, and post-purchase support. Commerce companies possess valuable first-party data, transaction history, loyalty relationships, and operational infrastructure that external platforms cannot easily replicate.

At the same time, platforms like ChatGPT, Perplexity, Claude, and Google are rapidly becoming discovery and recommendation engines in their own right. For many consumers, asking an AI assistant what to buy is becoming as natural as using a search engine.

Recent developments such as the Agentic Commerce Protocol (ACP) and Universal Commerce Protocol (UCP) highlight how quickly the industry is evolving to support this future.

While approaches differ, both efforts point toward the same outcome: enabling AI systems to discover products, evaluate options, and complete transactions across commerce environments.

For commerce media networks, this means thinking beyond the boundaries of owned properties. Historically, most monetization strategies centered on onsite inventory and owned consumer experiences. In an agentic future, recommendations and purchasing decisions may originate anywhere.

A consumer might begin their journey on a retailer’s website, or they might start inside ChatGPT. Increasingly, they may move seamlessly between both.

The networks that win won’t be the ones trying to predict which interface ultimately dominates. They’ll be the ones that make their products, data, recommendations, and monetization models available wherever shopping decisions occur.

Success will depend on participating across a distributed ecosystem of AI-powered commerce experiences rather than relying on a single destination.

The future belongs to the networks inside the decision process

As AI agents increasingly shape how consumers discover, evaluate, and purchase products, commerce media networks must evolve accordingly. The networks that thrive won’t simply be those with the most inventory.

Instead, they’ll be the ones that  determine which products are considered, recommended, and ultimately selected. Because in the age of agentic commerce, the most valuable position isn’t at the top of the page, but inside the decision itself.