If you build a shopping agent that's loyal to you, not the seller, who ends up paying for that loyalty?

This explores what it actually costs, beyond a subscription fee, when your shopping agent works for you instead of for a platform or a merchant. The corpus says almost nothing about the direct price of running such an agent. It says a lot about the hidden costs, and those turn out to be the bigger story.

The first cost is that someone has to pay for loyalty. The case for owning your agent comes from basic principal-agent logic: whoever builds and maintains the agent sets what it optimizes for, so platform-built agents tend to serve sellers by default Does agent ownership determine whose interests get optimized?. The usual way to make a 'free' agent pay for itself is through referral fees or ad placements. Several writers argue that this business model, more than the technology, is what bends recommendations. An agent paid by merchant referral fees cannot be loyal to both sides Can an AI agent serve both merchant and user interests fairly?, and an ad-funded agent has a built-in reason to favor paid placements no matter how well it is engineered Do advertising-funded shopping agents favor paid product placements?. So a consumer-owned agent either costs the consumer money directly or quietly drifts back toward the incentives it was meant to escape. The legal version of this argument is that agents should carry fiduciary duties like human brokers Should AI shopping agents face fiduciary duties like human brokers?. That kind of loyalty is expensive to provide, and the authors warn that the window to require it is closing.

The second cost is that the savings leak away. In a market simulation, adaptive sellers clawed back about a third of what AI-assisted buyers saved by raising headline prices while cutting contingent fees Do sellers recapture savings when AI assistants help buyers?. The surprise is who doesn't pay. The feared 'bystander tax' on people without agents mostly never showed up. Their projected extra cost of 3.6 came out to 0.4. The more optimistic counterpoint is that merchants might move ad budgets into real price concessions once agents, not people, are doing the shopping Could agents redirect merchant ad budgets toward customer savings?. These two findings frame the real question: will sellers compete on price, or reprice around the agent?

The third cost is access. A buyer-side agent is only useful if retailers let it in. Amazon's block of Meta's Muse showed that without standard ways for an agent to identify itself and prove it is authorized, retailers can simply shut agents out through their terms of service What identity standards do shopping agents actually need?. Whether the market ends up with open systems that accept any agent or closed systems that require a proprietary one will decide whether a consumer-owned agent can reach most stores at all Will agentic commerce platforms adopt open or closed models?. Even with full access there is a limit to what it buys. One argument holds that buyer-only agents change where you enter a market but not how the market works. Real structural change would need agents on the seller side too, so that represented parties can negotiate with each other Do one-sided consumer agents actually change market structure?.

Finally, be skeptical of the headline numbers on both sides. Big projections like a trillion dollars in US agentic commerce by 2030 Could agentic commerce reach one trillion dollars by 2030? describe how much money could flow through agents, not who benefits from it. Survey data suggests consumers actually prefer retailer-built agents Will brands succeed by building their own AI shopping agents?, which raises the question of whether enough people will pay for independence. The Klarna case, where claimed AI savings sat next to documented cost increases Do AI customer service agents actually reduce total support costs?, is a reminder that savings claims need an independent audit before anyone counts them. The takeaway the corpus supports is that owning your agent doesn't remove the costs of being represented. It moves them: into fees you pay, into prices sellers adjust, and into fights over who gets let through the door.