Retail Media Follows the Decision Surface

On September 10, 2026, one of the world’s largest e-commerce and advertising platforms confirmed a partnership with a leading conversational AI assistant to carry advertising inside its chat interface. Select advertisers in the United States can now extend their existing campaigns into a conversational environment, buying placements through the platform’s demand side offering without negotiating directly with the assistant’s operator. A global travel brand was named among the first testers.

The move positions retail media inside a surface that has moved from experiment to established business over the preceding year. It also signals a broader shift: as AI assistants become discovery surfaces, advertising is following.

The Mechanics

The pilot is limited to select advertisers and runs only in the United States. The platform framed the integration as a way for marketers to reach people who are actively using the assistant at the point where they explore options, compare choices and make decisions.

That framing is deliberate. The description situates the surface at the consideration and decision stages of the journey rather than at pure awareness. Travel is among the fastest growing verticals on the assistant through the summer of 2026, alongside financial services, software and health and wellness. A travel brand testing the surface first fits the pattern.

The buying model matters as much as the placement. Advertisers do not buy the inventory directly. They extend existing campaigns through an established demand side platform, which means budget, creative and measurement flow through systems that media buyers already use. In practical terms, conversational inventory has been added to an existing buying stack rather than treated as a separate channel requiring new workflows.

Why the Discovery Surface Is Moving

The strategic logic behind the move is straightforward. Traditional search is being disrupted. By 2026, traditional search volume was forecast to fall by as much as 25 percent as generative AI solutions became alternative answer engines for users. Consumers are increasingly starting product research inside a conversational interface rather than a search bar or a retailer’s website.

The traffic data reflects that shift. Referrals from generative AI sources to retail websites have grown sharply, with some measurement showing traffic from AI assistants to retail sites rising several times year over year through 2025 and into 2026. AI referred visitors also tend to arrive further along the funnel, with higher engagement and conversion rates than visitors from conventional search or social channels, because they arrive having already narrowed their options.

For retail media, the implication is structural. Retail media has been one of the fastest growing advertising segments for years, with global investment projected to reach $200.4 billion in 2026 and $223.4 billion in 2027, accounting for roughly 15 percent of total worldwide advertising investment. US omnichannel retail media alone is forecast to take in $71.67 billion in 2026. Yet almost all of that spend has been concentrated on retailer websites and apps, with in store media still representing less than 1 percent of the global market.

If the discovery moment is migrating into conversational interfaces, then retail media has to follow it. The alternative is to own the transaction while losing the decision.

What Remains Unsettled

Audience rules, inventory volume, measurement design and incremental outcomes are undisclosed. Attribution across the surface stays unsettled. This is a test, not a new retail media default.

Four questions will determine whether conversational advertising becomes a durable channel or a short lived experiment.

The first is incrementality. If advertising inside an assistant simply intercepts demand that would have converted through existing channels, it adds cost without adding sales. Answering that requires lift tests comparable to those now being introduced for in store media, where matched market frameworks compare sales at stores running campaigns against a control group.

The second is transparency. Users need to be able to distinguish an assistant’s genuine recommendation from a paid placement. Blurring that line risks the trust that makes the surface valuable in the first place, and trust is the asset that conversational interfaces cannot afford to lose.

The third is disclosure. Without published inventory volumes and audience rules, advertisers cannot plan and measurement partners cannot build comparable benchmarks.

The fourth is separation of assistance from influence. If the same system that advises a shopper also sells access to that advice, the commercial incentive and the user’s interest are no longer aligned. How platforms govern that tension will shape whether conversational advertising is accepted or rejected by the people it depends on.

The Retailer Imperative

For retailers and brands, the practical response is not to chase the pilot but to prepare for the pattern. Product data must be structured so that both AI assistants and advertising systems can interpret it consistently. Inventory signals must be live, because a recommendation that cannot be fulfilled is worse than no recommendation. Measurement frameworks must be rebuilt to isolate the effect of advertising inside AI mediated discovery rather than assuming it behaves like a search ad.

The brands that move first will not be the ones with the largest budgets. They will be the ones whose product information is clearest, whose availability is most reliable and whose measurement is honest enough to prove whether the channel works.

Where the Conversation Continues

The mechanics of conversational advertising can be described in a press release. The consequences cannot. Whether paid influence inside an AI assistant strengthens or erodes consumer trust, how incrementality should be measured when the discovery surface is a conversation and what retailers owe shoppers in transparency are questions that specifications and pilot programmes cannot answer on their own. They are resolved through debate, disagreement, and the exchange of experience between the people building these systems and the people running stores.

That conversation is happening in Paris this week. 

An Evening That Matters

At 7:00 PM tonight, September 16, the industry gathers for an exclusive networking cocktail and a Fireside Chat with some of the most influential voices in global retail. 

The discussion will address the forces reshaping the industry, from evolving consumer behaviour and the future of physical experiences to the adoption of new technologies across the value chain. It is an opportunity to hear directly from leaders who are navigating these shifts in real time, and to test your own thinking against peers facing the same questions.

The House of Innovation runs through tomorrow, September 17, in Hall 4 at Paris Expo Porte de Versailles. Whether you are exploring retail media, conversational commerce, agentic payments or the infrastructure that makes AI driven discovery work, this is where the technologies defining the next era of retail are being shown, tested and discussed. 

For full programme details, including tonight’s networking event and the final day of tours: Paris Connect

Request your invitation: Registration Form