Can a brand win over an AI shopping agent on merit alone, or does it all come down to who controls the platform?

This explores whether a brand can earn an AI shopping agent's favor on its own, or whether getting picked mostly comes down to who owns the agent and who runs the platform. The short answer from the collection: brands have less independent pull than they hope. When an agent picks something, it is reflecting the incentives of whoever built it and the hierarchies that already exist. A good product helps, but it rarely decides the outcome.

Start with the most uncomfortable finding. AI search engines cite brands almost entirely by how big they already are: 73% of the time for global names, 44% for mid-market brands, and 11% for niche ones. That pattern holds even when the user never names a brand in the query Do AI search engines favor well-known brands over others?. So 'becoming preferred' without platform leverage mostly works for brands that are already famous. Agents seem to amplify existing prominence more than they discover new winners.

The second force is ownership. The 'Coasean singularity' work argues that whoever builds and maintains an agent sets its goal. Platform-built agents tend to serve sellers, while agents the consumer controls can be tuned to the buyer Does agent ownership determine whose interests get optimized?. Ad-funded agents add another tilt toward paid placement, however well they are engineered Do advertising-funded shopping agents favor paid product placements?. Platforms and sellers can also hurt shoppers separately and still compound each other. In one simulation, a platform tuning its rankings and sellers tuning their product descriptions to the same agent blind spots more than doubled consumer harm, with no coordination between them Can platform ranking and seller descriptions separately harm AI shoppers?. A brand optimizing its descriptions without platform control is still playing on someone else's board.

There are cracks in that picture. One experiment found that adding a single sentence to an agent's prompt, naming the consumer as the person the recommendation serves, removed the platform-employment effect on how the agent treated sponsored products. Who employed the agent stayed the same Can naming the consumer fix AI agent sponsorship bias?. That suggests the bias is partly a matter of how the agent is instructed, not a fixed property of who owns it. Consumers also hold real power. In Accenture's survey, 56% said they would tell the agent which brands to buy, and 37% of loyal customers would not switch Do consumers actually want to delegate shopping decisions to AI?. For established brands, the most durable route to being 'preferred by the agent' may run through the human: the shopper names the brand, and the agent carries out the order. Accenture's strategic answer is for brands to build their own agents and become 'the agent of choice' Will brands succeed by building their own AI shopping agents?. Their evidence is survey sentiment, though, not proof that such agents work at scale.

What you might not have expected: the answer may depend less on brand strategy than on market structure. If agentic commerce splits into open systems, where many agents can operate, and closed ones that require a proprietary agent, much as payment networks split into open and closed loops, then 'without platform control' means very different things in each Will agentic commerce platforms adopt open or closed models?. A formal model adds that an algorithm's power to steer outcomes shrinks as more sellers compete Does algorithmic recommendation lose its edge as sellers compete more?. Crowded categories with many sellers may give brands more room than concentrated ones. Treat that as a theoretical hint. The collection doesn't yet have empirical work on smaller brands breaking through in agent-mediated markets.