If an AI shopper splits your grocery order across five stores to save money, what happens to the retailer that used to own that basket?

This explores which kinds of shopping are most exposed when an AI agent splits one shopping trip across several retailers, picking the best price for each item instead of leaving the whole order with one store. The corpus names only one category outright: grocery. Bain argues that agents able to build carts across several vendors, not just several products, could break up the retailer-owned basket. Competition would move from the store to the basket, with more price transparency and tighter margins as a result Could AI agents reassemble shopping baskets across multiple retailers?. Grocery fits the risk well. Baskets are large and repeat often, and many items are interchangeable, so an agent can compare them item by item without losing anything the shopper cares about. The collection has no ranked list beyond this. Everything else has to be pieced together from the conditions that make splitting a basket easy or hard.

The first condition is whether the purchase meets a need or creates a want. Rustamli argues that agents are good at fulfilling known needs, compressing search and purchase into a single zero-click step. They are poor at creating desire, which still comes from influencers and social platforms Do AI shopping agents destroy search ad revenue while leaving influencer marketing intact?. That suggests replenishment and commodity categories are the most exposed. Categories driven by taste and identity are probably less exposed, because the shopper arrives already wanting a specific thing. That is an inference from the corpus, not a measured result.

Travel is the useful counterexample, and it shows that fragmentation can be absorbed instead of destroying the incumbent. You might expect agents to scatter bookings across individual hotels and airlines. Goswami's data shows the reverse. Independent hotels' dependence on online travel agencies rose to 63.4%, because those agencies became the API and settlement layer that agents book through. Expedia's B2B business grew 24% Are AI agents actually removing online travel agencies from bookings?. In travel, whoever handles payment and fulfilment can end up owning the split basket. Whether this happens in grocery depends on whether someone builds the same kind of settlement layer there.

The second condition is who owns the agent and how open the platform is. Bain separates third-party agents like ChatGPT from retailers' own on-site agents like Amazon Rufus. Shoppers trust the retailer's agent three times more, and on-site agents drove $12B in extra sales How are different AI agents reshaping the retail shopper journey?. A retailer's own agent keeps the basket in one place. Fragmentation requires a cross-retailer agent working on an open system rather than a closed one Will agentic commerce platforms adopt open or closed models?. It also requires the agent to be able to carry your purchase history and preferences between stores, which the UK's competition regulator treats as basic market infrastructure Should data portability be infrastructure for agent-driven markets?. Choudary adds a limit: agents that work only for the buyer change where shoppers enter a market, not how prices are set. Deeper change needs agents on the seller side too Do one-sided consumer agents actually change market structure?.

The takeaway: the most vulnerable categories combine large repeat baskets, interchangeable products, need-driven buying and weak loyalty to a retailer's own agent, and grocery is the clearest case. Fragmentation does not always hurt incumbents, though. As travel shows, the winner can be whoever runs the plumbing the agents use. That fits BCG's finding that consumers and platform networks gain most from agent shopping while individual merchants carry the structural disadvantage Who wins and loses as AI agents reshape shopping?.