As agentic commerce reshapes retail, it forces brands to rethink everything from product discovery to purchase completion.

"The biggest threat of all of this agentic commerce is the impact it has on how people discover new products to buy, how they find new things, and make purchase decisions," said Jason Goldberg, chief commerce strategy officer at Publicis Groupe, during a recent Publicis Commerce webinar. "This is the disruption of discovery."

Building for the agentic shelf

Winning in agentic commerce requires a shift in operating models. Four pillars determine success, said Julia Miller, group vice president of commerce media for Mars United and Publicis Commerce.

• Retail readiness: In-stock availability, competitive pricing, and strong ratings and reviews remain foundational. These trust signals are non-negotiable for agent recommendations.
• Content optimization: Product detail pages must serve both human and robot audiences. While humans respond to visual lifestyle imagery, robots prefer detailed text including 200-page PDFs with comprehensive specifications, dietary attributes, and usage occasions.
• Catalog enhancement: Brands must populate enhanced, contextual attribute fields that enable products to appear in lifestyle and occasion-based searches.
• Strategic media investment: Paid placement matters, but cannot substitute for trust. Sponsored results must still meet relevance criteria or risk eroding platform credibility.

"You'll need five times more data to support yourself [in agentic commerce] than you did on the digital shelf,” said Miller, citing data from a recent Adobe study.

Context wins over blanket optimization

The recommendation algorithms in agentic environments respond to specificity, not volume.

"The question is no longer just, how do I show up more? It's how do I become the right answer for this very specific shopper moment," said Tyler Rosten, director of ecommerce at Saatchi X.

In testing with a health and beauty brand on Walmart's Sparky platform, Rosten's team discovered that certain prompts instituted price thresholds that excluded premium products entirely. However, different prompts for the same product showed strong results after content optimization, with price being less of a factor.

Brands cannot optimize every product for every query. Success requires identifying which products fit which shopper missions, then optimizing strategically for those specific contexts.

"This is not about winning every prompt," Rosten said. "It's about deciding which products are relevant for the queries, and then optimizing the PDPs to win."

Product detail pages (PDPs) must become dynamic rather than static, rotating content based on seasons, occasions, and trending topics. Brands treating PDPs as "set and forget" assets will lose to competitors who update content quarterly or in response to cultural moments.

The formula for AI recommendation visibility

"The real estate for brands in AI search is considerably smaller than in traditional retail environments," said Ethan Goodman, president of Profitero Plus. "On average, eight items are recommended by an AI agent during a search, making optimization critical to earning a place in the consideration set."

Proprietary research from Profitero, Mars United Commerce, and Publicis Commerce studying Amazon's agent (then Rufus, now Alexa Plus) revealed a two-stage model for AI visibility: eligibility and optimization.

First, products should meet baseline criteria including bestseller status (ranking in top 15,000), 4.6+ star ratings, and substantial review volume. Without meeting this threshold, content optimization has minimal impact, according to Goodman.

Once products clear the eligibility bar, strategic content updates can significantly improve "share of agent recommendations" (SOAR), or the percentage of times a product appears in AI-generated results.

"There's a certain threshold for eligibility that brands and products need to meet in order to be recommended," Goodman said. "Once brands have met that threshold, content optimization can influence where they show up within those recommendations."

This was originally featured in the Retail Daily newsletter. For more retail insights, statistics, and trends, subscribe here.