Major retailer extends sponsored products into AI shopping experiences, powered by Koddi

Koddi enables a top retailer to bring its existing sponsored product offering into new conversational shopping experiences

As shoppers increasingly turn to AI-powered experiences for discovery, planning, and purchasing, the conversation itself is becoming a new commerce media surface. More than three-quarters of advertisers say they are likely to invest in agentic commerce models, creating an opportunity for commerce media networks to bring existing advertiser demand into these emerging journeys.

A leading retailer was developing an AI shopping tool designed to help customers accomplish more complex tasks, from planning meals to finding products that meet specific needs. The retailer saw an opportunity to extend its established commerce media program into this experience, giving advertisers another way to reach shoppers during moments of consideration and purchase. 

The challenge was doing so without compromising the usefulness of the assistant. The result: a powerful new format now achieving 100% fill rate and scaling intelligently across thousands of conversations each day.

The challenge: Turn a conversation into an ad opportunity

Traditional sponsored search starts with a relatively structured signal. A shopper searches for an item, then the ad server identifies eligible products, runs an auction, and returns a sponsored result.

Conversational shopping is different. There may be no explicit product search, yet the conversation contains several signals that can inform relevant recommendations.

To introduce sponsored products into the experience, the retailer needed a way to:

• Interpret broad, goal-oriented shopper requests
• Translate conversational context into signals its existing ad server could understand
• Determine when a conversation represented an appropriate advertising opportunity
• Identify sponsored products that genuinely satisfied the shopper’s needs
• Run an auction without adding unnecessary friction or latency
• Return the winning sponsored product to the AI experience naturally and transparently

The goal was to make conversational intent actionable within the retailer’s existing commerce media infrastructure, doubling down on their ability to plan, manage, and report on all media in one unified platform.

The solution: Connect conversational intent to the existing auction

Koddi worked with the retailer to enable its AI experience to call into its existing sponsored product offering.

Rather than treating the shopper’s latest message as a traditional search query, the experience translates relevant context from the conversation into structured commerce signals.

Those signals can include product categories, relevant products or SKUs, attributes, inventory requirements, location, and other constraints derived from what the shopper has said. That connection depends on structured, AI-ready product data. It is a priority advertisers recognize as well: 86% say they plan to invest in data integrations and AI-ready feeds for agentic commerce. For commerce media networks, those feeds become the bridge between what a shopper tells an AI assistant and what the advertising system can actually serve.

Koddi then uses those signals to determine eligible sponsored products and run the same core auction logic that powers the retailer’s broader commerce media program.

The result is a flow that connects the new consumer experience to established advertising infrastructure:

Shopper conversation → intent and product needs → eligible sponsored products → Koddi auction → sponsored recommendation → AI response

This gives the retailer a path to monetize a new AI surface while allowing advertisers to participate through the commerce media ecosystem they already use.

Preserving contextual relevancy and existing auction logic

A recommendation from an AI assistant can carry more weight than a conventional ad placement. That makes relevance especially important, so Koddi has to inform two key decisions.

First, the system determines whether a sponsored product satisfies the shopper’s request. Product taxonomy, attributes, inventory, location, fulfillment requirements, and conversational constraints can all be used to establish eligibility. Only products that meet the appropriate relevance threshold move forward.

Then Koddi applies commercial auction signals such as bid, Quality Score, predicted performance, and advertiser eligibility to rank the remaining sponsored candidates.

That distinction enables the retailer to preserve the usefulness of its AI experience while still creating meaningful opportunities for advertisers.

Built for conversations, not keywords

Conversational commerce also introduces another important difference: shopper intent evolves.

A customer might begin by asking for dinner ideas, later mention a dietary restriction, and finally ask for everything to be added to their cart. Each turn adds information that can change which products are relevant and which sponsored products should be eligible.

The integration is designed to account for that accumulated intent rather than evaluating each message in isolation.

That enables advertising to become more responsive to the shopper’s actual journey through the conversation, including changing preferences, requirements, and purchase intent.

Extending commerce media into a new shopping surface

With Koddi powering the ad-serving layer, the retailer was able to extend sponsored product advertising into its AI shopping experience while continuing to leverage its existing advertiser demand, product data, auction mechanics, and commerce media infrastructure. Advertisers can simply log into their existing self-serve portal to activate these new campaigns. 

For advertisers, the new experience creates an opportunity to participate in a different kind of shopping moment. Instead of appearing only after a shopper searches for a specific product, brands can become eligible when their products are relevant to the broader goal a shopper is trying to accomplish.

For the retailer, conversational advertising creates a path to monetize emerging AI experiences without treating them as a separate media business.

And for shoppers, the experience can continue to prioritize the products that best satisfy what they are trying to accomplish, with sponsored influence incorporated transparently into the decision.

The bigger opportunity

Traditional search asks commerce media systems to understand what someone typed, but conversational shopping asks them to understand what someone is trying to accomplish.

That shift requires new decisioning between the conversation and the auction, but it does not require abandoning the commerce media infrastructure networks have already built. By connecting rich conversational context with structured product data, relevance controls, and existing auction mechanics, commerce media networks can make AI shopping experiences another addressable surface for advertisers while protecting the consumer experience that makes those assistants valuable in the first place.

The opportunity also extends beyond lower-funnel product searches. Sixty-four percent of advertisers expect their largest agentic commerce investments to focus on mid-funnel consideration and research, where shoppers are asking agents to compare options, solve problems, and build consideration sets. Conversational ad serving gives commerce media networks a way to make those moments addressable while maintaining relevance as the first requirement for participation.