Key takeaway
Agentic AI is beginning to change not only how consumers discover products, but how purchasing decisions are made. As AI systems compare products, recommend suppliers and complete transactions on a user’s behalf, brand owners face a different set of legal and commercial risks: counterfeit products may enter automated recommendation pathways, misleading product claims may influence machine-led purchasing, and traditional assumptions about consumer confusion may become harder to apply where software intermediates the transaction.
For legal and commercial teams, the question is whether existing trade mark, advertising, consumer protection and enforcement strategies are equipped for a market in which software decides what the customer sees and buys.
What is agentic AI shopping?
Agentic AI shopping refers to the use of AI systems that can search for products, compare options, make recommendations and, in some cases, complete purchases on a consumer’s behalf with limited further input.
A consumer asks an AI assistant to find a waterproof running shoe under £120, compare suitable options, select a size and place the order. The consumer sets the objective, but the AI system carries out much of the product discovery, comparison and selection process itself.
This differs from traditional e-commerce, where the consumer typically searches, browses product pages, compares alternatives and makes the final purchasing decision directly.
In an agentic shopping journey, the AI system becomes an intermediary between the consumer and the market. It determines which products are considered, which information is given weight and which supplier is ultimately selected.
For brand owners, that means product discovery and purchasing decisions are influenced by software before the consumer ever sees the competing options.
Why does agentic AI matter for brand protection?
Much of trade mark and brand protection law has developed around what human consumers see and understand.
Where competing products use similar names, packaging or branding, the legal analysis may consider whether consumers are likely to be confused about origin or commercial connection.
Agentic purchasing complicates that picture.
If an AI system recommends or purchases a look-alike product without the consumer seeing the competing listings, the human may not encounter the potentially infringing branding until after the transaction has already taken place.
The fact that software intermediated the purchase does not make trade mark rights disappear.
But it may affect the factual questions that arise when assessing how the product reached the customer, what representations influenced the purchase and where any confusion occurred.
That will matter particularly where AI shopping systems begin to move beyond recommendation into autonomous purchasing.
Brand owners should therefore expect enforcement disputes to involve evidence about how products are represented to AI systems as well as to human consumers.
For businesses expanding or investing materially in their brands, clear ownership of registered rights remains fundamental. See our article on trade mark protection for business expansion and whether you own your brand.
What happens when an AI agent recommends a counterfeit product?
Counterfeit products already present a substantial enforcement problem across online marketplaces and social platforms.
Agentic shopping introduces another route through which those products may reach consumers.
A counterfeit seller may not need to persuade the customer directly. It may only need to appear sufficiently relevant or credible to the AI system conducting the search.
That creates obvious opportunities for manipulation.
A listing might mimic the language used by an established brand, reproduce product descriptions, use misleading specifications or position itself as a close alternative to a genuine product.
If an AI agent treats those signals as reliable, the counterfeit product may be recommended or purchased before the consumer has ever seen it. Brand owners will need to understand whether those listings are also entering AI-generated recommendations.
That could make evidence about product feeds and recommendation outputs crucial in future infringement disputes.
Can AI shopping change how consumer confusion is assessed?
Traditional trade mark analysis often focuses on the perception of the average consumer. Agentic purchasing raises a more complicated question: what happens where the human consumer does not make the comparison that led to the purchase?
The AI system itself is not a legal consumer capable of being confused in the conventional sense. But the transaction still happens for the benefit of a human user whose purchasing decision has effectively been delegated.
That does not require an entirely new legal framework overnight. Existing principles may remain capable of dealing with many disputes.
But the evidence will look different. Questions could include:
- what information the AI system relied on;
- which products it considered;
- how competing goods were described;
- whether the consumer specified a particular brand;
- whether the agent substituted a competing or look-alike product; and
- what the consumer saw before and after the transaction.
Could misleading product claims become more important in agentic commerce?
AI shopping systems rely heavily on information that can be compared quickly: specifications, pricing, availability, review scores, delivery terms and product claims.
Naturally, this creates a strong commercial incentive for businesses to present their products in a way that AI systems can easily identify and rank.
It also creates legal risk.
A claim does not become permissible merely because it is designed to influence an algorithm rather than a consumer directly. Comparative claims, performance statements and product descriptions still need to be accurate and capable of substantiation.
The same applies where businesses create content designed to position their product as superior to a named competitor. If agentic commerce increases the value of machine-readable marketing claims, legal teams will need to exercise greater oversight over how those claims are produced and distributed.
The pressure to make information attractive to AI systems should not result in a lower standard of legal review. We discuss a related issue in our article on AI-washing, misleading AI claims and litigation exposure.
Could competitors manipulate AI systems using your brand?
One concern for established brands is that competitors or counterfeit sellers may seek to benefit from the signals associated with a recognised brand without necessarily copying it in the most obvious way.
They might reproduce descriptive language associated with the brand, create comparison content designed to intercept relevant queries or structure listings in a way intended to appear as an equivalent or substitute product.
This may not always amount to trade mark infringement, but depending on the conduct, it could raise questions around trade mark use, passing off, misleading advertising or unfair commercial practices.
That means legal teams may eventually need to monitor a much wider range of conduct than traditional keyword advertising or marketplace listings.
The relevant question will become not only “what appears when a customer searches for our brand?”, but also “what does an AI system recommend when a customer describes the need our brand is known for solving?”
Should brands monitor what AI systems say about them?
Incorrect or misleading outputs can have commercial consequences even where the business itself did not publish them.
Monitoring can help identify:
- counterfeit or look-alike products appearing in recommendations;
- incorrect descriptions of the brand;
- false comparisons with competitors;
- outdated product information;
- confusion between genuine and unauthorised sellers; and
- claims being attributed to the business that it has never made.
The legal remedies available will depend heavily on where the underlying information originated. An AI-generated statement may simply be reproducing information published elsewhere, in which case the real issue may lie with the original seller or publisher rather than the AI provider itself.
Tracing the source of the output will become a huge part of enforcement strategy.
How should brand owners prepare for agentic commerce?
The immediate response does not require an entirely new legal strategy. It requires existing brand, advertising and commercial controls to be applied to a new distribution environment.
Brand owners should consider:
Protecting core trade marks. Registered rights remain one of the strongest foundations for taking action against counterfeit and confusingly similar products.
Improving control of product information. Product names, specifications, authorised seller details and key claims should be accurate and consistent across relevant channels.
Reviewing comparative marketing. Content designed to improve visibility in AI recommendations should be subject to the same legal scrutiny as conventional advertising.
Monitoring new recommendation channels. Businesses may need to test how major AI systems describe and recommend their products alongside existing marketplace monitoring.
Preserving evidence. Where an AI system recommends an infringing or counterfeit product, screenshots, prompts, outputs, product listings and transaction details may all become relevant evidence.
Reviewing distribution arrangements. Clear authorised reseller and distribution structures are crucial where AI systems need to distinguish genuine sellers from unauthorised ones.
Coordinating legal and marketing teams. Product-data optimisation and AI visibility should not develop entirely separately from brand protection and advertising compliance.
What does agentic AI mean for the future of brand enforcement?
Agentic commerce will transform the environment in which familiar rights need to be enforced.
Trade marks will still identify origin. Counterfeit products will still infringe rights. Misleading commercial claims will still create exposure.
But the route between the offending conduct and the consumer will include an AI system making decisions on the consumer’s behalf, which means brand owners will need to think beyond what customers themselves see.
The next generation of brand protection is going to involve understanding what information AI systems can access, which products they recommend and how bad actors may try to influence those recommendations.
The core challenge is whether enforcement and compliance practices can adapt quickly enough to protect the brand when the customer is no longer the only one making the purchasing decision.
Barnes Law’s Intellectual Property team advises businesses on trade marks, brand protection, infringement, licensing and disputes involving the misuse of intellectual property. For more information, please contact our Intellectual Property team to discuss how we can support you. Our Intellectual Property Disputes team advises on enforcement strategy, injunctions and proceedings before the High Court, IPEC and UKIPO.
Written by Barnes Law Managing Partner Yulia Barnes.
