The advent of agentic commerce, where artificial intelligence (AI) systems autonomously evaluate, recommend, and even purchase products on behalf of consumers, represents a profound shift in the landscape of product discovery and brand interaction. As these AI agents become more sophisticated and integrated into daily life, the traditional marketing focus on brand building and direct customer engagement is being fundamentally challenged. A brand chosen by an AI algorithm for its legibility and optimized data points is a brand procured, not necessarily preferred. As strategist Jess Graham aptly notes, this is akin to a procurement process, and genuine human connection or loyalty rarely blossoms from mere sourcing.
This distinction holds significant weight as AI agents assume an increasingly central role in the initial stages of product discovery. Brands meticulously designed to be ‘readable’ by machines, with structured data and optimized keywords, risk becoming ‘invisible’ to the human behind the agent. The AI performs its evaluation, the customer receives a product, and the critical act of conscious human choice subtly evaporates from the transaction. For marketing professionals responsible for martech infrastructure and customer data, this emergent paradigm presents a formidable strategic test: balancing the undeniable necessity of algorithmic legibility with the enduring imperative of maintaining a meaningful presence in the human consciousness that ultimately drives repeat purchases and brand loyalty.
Early Forays and Pivots: A Chronology of Agentic Commerce
The concept of agentic commerce is not merely theoretical; it is already manifesting in the market, albeit in a more nuanced form than initially envisioned by early headlines. The rapid evolution of generative AI capabilities, particularly following the widespread public access to tools like ChatGPT in late 2022, spurred a flurry of innovation in how AI could facilitate commercial transactions.
One of the most notable early initiatives was OpenAI’s Instant Checkout, launched in September 2025. This ambitious feature aimed to integrate a seamless purchasing experience directly within the ChatGPT interface. However, its lifespan proved short-lived; it was discontinued in March 2026, just five months after its debut. The reasons for its withdrawal highlighted critical insights into consumer behavior and merchant readiness. Only a limited number of merchants, approximately a dozen, had successfully integrated the feature. Furthermore, usage remained low, indicating a significant disconnect between the AI’s ability to recommend and the consumer’s willingness to complete a purchase within a third-party AI environment. Shoppers who utilized ChatGPT for product research largely preferred to finalize their transactions on the familiar and trusted websites of the retailers themselves.
OpenAI did not abandon the commerce space entirely but strategically pivoted its approach. Recognizing the consumer preference for retailer-owned checkout experiences, it transitioned to a discovery-first model. This revised strategy routes shoppers from ChatGPT directly to merchant applications and storefronts, where the retailers retain control over the payment processing and, crucially, the customer relationship. This pivot underscored the complexity of integrating AI into the full commerce funnel, particularly the sensitive checkout layer.
In contrast, Google pursued a different, more integrated protocol with its Universal Commerce Protocol (UCP). Announced at the National Retail Federation (NRF) conference in January 2026, the UCP represented a collaborative effort with major players including Shopify, Etsy, Wayfair, Target, Walmart, and Visa. This protocol aimed to establish a standardized framework for AI agents to interact with retail ecosystems. It has since gone live and continues to expand its functionalities, incorporating features such as cart support, comprehensive catalog access, and identity linking. These capabilities enable AI agents to perform more sophisticated actions on behalf of users, facilitating a deeper integration into the discovery and pre-purchase phases of the shopping journey.
Comparing the trajectories of OpenAI and Google reveals a clear trend: while the optimal configuration for the checkout layer remains an area of ongoing experimentation and evolution, AI agents have firmly established their presence in the discovery layer. This front-end integration is where the immediate battle for brand visibility and customer influence is being waged.
The Rise of AI in Product Discovery: Data and Trends
The influence of AI on product discovery is not merely anecdotal; it is increasingly substantiated by consumer research. A 2025 study by Salesforce revealed that a significant 39% of consumers have already utilized generative AI tools to discover and evaluate products. This trend is even more pronounced among younger demographics, with 54% of Gen Z consumers reporting similar usage. This demographic, often early adopters of new technologies, signals a future where AI-driven discovery will become an ingrained aspect of the shopping experience.
Adobe’s tracking data further corroborates this trend, indicating a measurable increase in traffic directed from generative AI tools to retail websites. This suggests that AI is effectively serving as an intermediary, guiding consumers from initial query to potential purchase points. However, the data also highlights a crucial nuance: while consumers are leveraging AI for discovery, their willingness to allow AI agents to make purchases on their behalf is still nascent. Salesforce found that 63% of Gen Z consumers express interest in having AI agents execute purchases for them. While this indicates a strong future potential, it reflects stated interest rather than actual, measured behavior. The gap between interest and adoption underscores a lingering need for trust, control, and perhaps, a more compelling value proposition for full agentic purchasing.
Consider a near-future scenario, vividly painted by Jess Graham, where a customer instructs an AI agent to purchase a moisturizer. The agent, leveraging its algorithms, instantaneously sifts through thousands of options, meticulously weighing factors such as price, user reviews, ingredient lists, and delivery speed. It then selects and purchases a product. In this transaction, the customer never sees the packaging, never engages with the brand narrative, and never has the opportunity to compare it against a product they might have previously used and loved. The agent makes the choice, and the customer receives the outcome, effectively bypassing the traditional brand touchpoints that foster loyalty.
The Peril of Agentic Invisibility: Brand Erosion and the Discovery Tax
At the heart of the challenge for brands lies the critical distinction between algorithmic legibility and genuine brand preference, as articulated by Graham. Algorithmic legibility refers to the technical optimization of product data – ensuring it is structured, tagged, and presented in a manner that AI agents can easily read, categorize, rank, and ultimately include in a consideration set. This pursuit, often termed "top of algorithm," has seen significant budget allocation from brands, yielding tangible early returns. Research from Seer Interactive in September 2025 indicated that brands appearing in AI-generated answers experienced increased traffic. Furthermore, Adobe’s data confirmed that this AI-driven traffic tends to result in longer site visits, suggesting higher engagement. This work is undoubtedly important and must continue as a baseline operational requirement.
However, the more profound challenge is maintaining a meaningful presence with the human consumer who ultimately benefits from the AI’s recommendations. The failure state in this scenario is "agentic invisibility," a condition where a brand, despite being algorithmically legible, becomes indistinguishable or irrelevant to the human shopper. The commercial implications of this invisibility are severe and direct. A brand purchased without a deliberate human choice risks surrendering its pricing power.
To an AI agent, constantly optimizing for objective metrics like price, aggregated ratings, and delivery efficiency, an undifferentiated brand appears as a mere commodity. In a commoditized market, competition invariably devolves into a race to the bottom on price, where profit margins erode, and long-term brand equity suffers. Graham terms the compounding cost of ceding control over how customers discover and choose a brand as the "discovery tax." This tax, which directly impacts a brand’s profit and loss statement, represents the ongoing erosion of value when brands fail to cultivate a distinct preference that transcends mere algorithmic selection. It is the cost of being chosen by a machine, but not truly loved by a person.
Lessons from History: Industries That Ceded Discovery
The pattern of industries relinquishing control over discovery and subsequently paying a significant price is not unprecedented. Numerous sectors have navigated similar shifts, often with stark consequences for brand differentiation and profitability.
Consider the travel industry. Before the widespread adoption of online travel agencies (OTAs) like Expedia and Booking.com, airlines, hotels, and car rental companies maintained direct relationships with their customers through their own booking channels and traditional travel agents. OTAs offered unparalleled convenience, aggregating options and simplifying the booking process. To maintain access to a vast customer base, brands increasingly relied on these intermediaries, accepting worse economics (e.g., commissions, reduced control over pricing and promotions). Over time, OTAs captured the customer relationship, amassed valuable data on preferences and booking patterns, and effectively commoditized travel products. Hotels, for instance, often became interchangeable features within an OTA search result, differentiated primarily by price and star ratings rather than unique brand experiences.
A similar trajectory can be observed in the early days of e-commerce for many product categories. Brands that relied heavily on large online marketplaces for distribution often found themselves competing primarily on price and reviews within a platform that owned the customer data and relationship. The marketplace dictated the terms, and individual brands struggled to cultivate direct loyalty outside of the platform’s ecosystem.
The sequence repeats: a new intermediary, offering unprecedented convenience (in this case, AI agents), emerges. Brands, eager to maintain access to evolving customer channels, accept increasingly unfavorable economics. The intermediary then captures the vital customer relationship and proprietary data, leading to a collapse of brand differentiation into a mere feature comparison. Agentic commerce represents the most complete manifestation of this pattern to date, precisely because the human customer is not even actively present during the critical evaluation phase. The choice is made for them, by an algorithm.
Reclaiming Control: The Imperative of First-Party Data
Graham’s prescription for brands facing the agentic shift is clear: proactively build discovery experiences that AI agents cannot fully capture, thereby giving people compelling reasons to choose a brand from the outset. The martech stack plays a pivotal role in this strategy, as earning genuine preference demands more than simply making a brand legible to an algorithm.
Many brands currently chasing algorithmic legibility may inadvertently be feeding AI agents the wrong kind of data. Static product feeds and last-touch attribution models, while useful for describing what a customer did, remain silent on the more critical question of why they chose a particular brand. This level of data is sufficient for algorithmic ranking but falls short of fostering true brand preference. An AI agent, optimized to weigh a few structured fields, will inevitably treat every well-structured competitor as interchangeable.
This inherent instability is demonstrably evident in current AI results. SparkToro’s research revealed a striking finding: there was less than a 1% chance that the same brand would appear across two identical AI queries. A brand might flicker into prominence in one answer, only to vanish without explanation from the next, leaving both marketers and customers in the dark. This unpredictability underscores the precariousness of relying solely on algorithmic legibility.
Escaping this state of agentic invisibility thus becomes a data architecture decision before it translates into a campaign strategy. It necessitates the collection and strategic utilization of zero- and first-party data that captures nuanced preference and relationship signals – the qualitative insights that explain why a customer chose a brand, and crucially, why they might choose it again. This invaluable data must be securely housed within a brand’s own martech stack, under its direct control.
In practical terms, this means brands must proactively own the discovery moment rather than merely borrowing it through third-party data or walled-garden platforms. It involves cultivating and capturing signals that emerge from direct community engagement, one-on-one conversations, and even treating the post-purchase experience (delivery, unboxing, customer service interactions) as rich sources of relationship data, not just logistics. Furthermore, building a consented identity framework that persists regardless of whether an AI agent mediates the interaction is paramount. A brand that successfully owns its discovery journey, cultivates direct customer relationships, and controls the underlying data, positions itself to become the brand an AI agent can be specifically instructed to seek by name. This represents the durable, resilient version of "top of algorithm," built on data that the brand itself governs and leverages for enduring preference.
Strategic Imperatives for Marketers: Architecture, Audit, and Ownership
From a data perspective, the stark reality is that addressing agentic invisibility is fundamentally an architectural challenge before it becomes a strategic one. A marketing team struggling to connect its existing martech tools will find it exceedingly difficult to suddenly capture and leverage granular customer preferences and relationship signals. The emergence of agentic commerce doesn’t create this data gap; rather, it brutally exposes and amplifies its consequences, simultaneously raising the cost of inaction.
Marketers must initiate this transformation with a comprehensive data audit, rather than immediately seeking new tools. The first step is to meticulously map the zero- and first-party customer data currently being captured across all touchpoints. This audit demands brutal honesty: a dataset that merely records what was bought, with minimal insight into why it was chosen, is designed purely for algorithmic legibility and offers little beyond that.
Following this audit, a deliberate decision must be made regarding where vital relationship data will be built and securely stored – exclusively within the brand’s own ecosystem. This includes investing in owned discovery channels, fostering vibrant customer communities, optimizing the delivery and unboxing experience for relationship building (as highlighted by Graham), and establishing direct communication channels that are impervious to platform freeze-outs. The focus must shift from merely optimizing for the machine to strategically investing beyond it, ensuring algorithmic legibility is treated as the absolute minimum requirement.
Furthermore, marketers should quantify the "discovery tax" in tangible financial terms for their organizations. Presenting a board with the stark reality that "we win the transaction but lose the relationship" is far more likely to secure funding for necessary fixes than simply delivering another dashboard of vanity metrics. This financial articulation underscores the long-term impact on brand equity and profitability.
Finally, the ownership of agent-facing data must reside where the customer relationship inherently lives: within the marketing department. Allowing this critical function to drift into a purely technical decision made downstream by IT or product teams risks detaching it from the core strategic objectives of brand building and customer engagement.
Agentic commerce is poised to hand transactions to the brands most algorithmically legible to machines. However, whether it also relinquishes the invaluable customer relationship is a strategic martech and data decision being made at this very moment. Marketers must seize the initiative and lead this decision, rather than allowing it to be dictated by default, thereby safeguarding brand value in an increasingly autonomous commercial landscape.
Related posts:
