With agentic commerce, brands will have to convince their customers, but also the AI which will guide, compare and arbitrate their purchases, placing customer data and CRM at the center of the game.
Retail is moving into the era of agentic commerce, where the act of purchasing will soon be delegated to autonomous AI. The competition is no longer played exclusively on the customer journey, but on the arbitration of algorithms. Here’s how brands must reinvent themselves to exist in this new ecosystem of automated fulfillment.
We are witnessing a historic paradigm shift: the era of web consultation is giving way to that of autonomous execution. AI no longer just answers our questions: with ChatGPT’s Instant Checkout, Alexa for Shopping at Amazon or new shopping experiences from Google (AI Mode, virtual fitting and agentic checkout), it takes action. This shift is not theoretical, it is already visible in several markets where search and shopping experiences augmented by AI are deployed, notably in Spain, Italy and Portugal. The consumer no longer browses, he delegates. The search and comparison of offers are now entrusted to agents capable, in certain cases, of deciding and paying on their behalf.
For retailers, this turning point is not a simple technological adjustment, but a profound restructuring of the purchasing journey. Retailers must prepare now, because what is happening today in some markets is already redefining the conditions for visibility, preference and conversion tomorrow. A strategic question then arises: how can we exist when the interface (website, application, etc.) is erased in favor of algorithmic arbitration?
From awareness to “cognitive preference”
In this context where direct interaction is becoming rarer, the battle is no longer just about visibility, but about the ability to be prioritized ahead of any interaction. The differentiation must be formulated explicitly to become an interpretable signal. If the AI agent is programmed to optimize the purchase, it must also be able to identify certain values, beyond the simple notion of price: ethics, sustainability, excellence.
The goal is to anchor a cognitive preference: that the user tells their AI to choose a specific brand rather than a product category. Without this strong identity, the retailer risks disappearing behind a gigantic flow of data, reduced to a simple interchangeable logistics warehouse.
Semantic marketing: nourishing knowledge graphs
For a brand to be effectively prioritized by an agent, its value proposition must be readable and usable by the systems that arbitrate. In other words, the brand promise is only valid if it is encoded. To be chosen, each product must become an exhaustive digital entity, integrated into knowledge graphs via protocols such as MCP (Model Context Protocol).
Each feature, each label and each competitive advantage must be structured so that the agent can verify the digital “proof” of it. Visibility no longer depends on the aesthetics of a site, but on the ability to make value readable and immediately comparable in automated environments.
Loyalty as a real-time negotiation protocol
Once this readability is ensured, the question becomes that of the lasting relationship between the brand and the customer in an environment mediated by agents. This is the reason why loyalty also changes its nature. It moves from a catalog of benefits for humans to a system of benefits exposed by API. For the preference to be sustainable, the associated advantages (discounts, delivery priorities, exclusivities, etc.) must be identifiable by the agent at the time of arbitration.
Loyalty thus becomes a dynamic negotiation protocol: if the agent cannot “see” the immediate gain for its user, it will direct the act of purchase towards a more generic alternative.
CRM, knowledge and context infrastructure
A true pivot of the agentic strategy, the CRM is transformed into a reservoir of actionable contexts which makes it possible to precisely align the offer with the buyer’s intention. By securely exposing usage histories, size preferences or repurchase cycles, the retailer provides the AI agent with unprecedented discernment capacity to refine its requests.
From then on, customer knowledge is no longer content with documenting the past: it becomes a predictive and contextual lever allowing AI to anticipate demand and guide the act of purchasing upstream, well before friction appears.
Making trust a certified digital asset
Even a well-identified brand will only be retained if it offers guarantees of reliability that can be interpreted by agents, making trust a competitive advantage once its commitments are not only displayed, but certified and technically usable by machines. This is where standards like UCP (Universal Commerce Protocol) come in, which secure transactional exchanges between merchants and agents.
In the age of autonomous commerce, reliability is a signal of performance. An agent will without hesitation ban a merchant whose deadlines are uncertain or whose return policy is opaque. Trust now translates into proof of technical performance, turning each successful transaction into a positive score for future algorithms.
In conclusion, agentic commerce does not signify the end of the customer relationship, but its profound change. To win, retailers will have to agree to become “readable” by machines while remaining “indispensable” to humans. The battle will no longer be won on the click, but on the ability to become priority data and a reflex of trust in the invisible ecosystem of personal agents.
