Today’s consumers don’t start shopping with a list; they start with a question. “What should I buy for my teenage nephew who likes anime and cycling?” “What’s a thoughtful gift for grandparents who live far away?” “Can you recommend fun, office-appropriate presents under £10?” That shift is already visible in real purchasing behavior: last Christmas, Santa took a back seat. AI decided what gifts we unwrapped, and they were bought from the retailers that could automatically and accurately answer those customer questions.
Online retailers can’t simply drop a chatbot into their website or app and hope for the best. Everything from product listings to purchasing patterns and customer reviews must be meticulously structured for easy retrieval by AI systems, including the generative search engines that are fast becoming the first stop in shopping journeys. And with ChatGPT rolling out commerce integrations directly into its interface, understanding and leveraging AI’s position in the funnel will only become more vital to sales success.
The Fast-Approaching Query-Based Purchase Funnel
Generative platforms are training consumers to pick products via prompts rather than endless scrolling through listings. While AI-assisted shopping accounts for only a relatively small share of overall traffic, habits are changing rapidly. Between January and July of 2024, Adobe found that AI traffic to U.S. retailers more than quadrupled. The transition to a conversational sales funnel is picking up pace, and the first impression a chatbot provides for a product will matter as much as any listing or homepage.
In the U.S., this transition is moving even faster thanks to emerging in-chat purchasing systems, such as Instant Checkout within ChatGPT, which enables single-item purchases from merchants on selected platforms. Even when a customer buys directly from a retailer’s site, there’s a high chance the recommendation that nudged them came from AI.
Advertisement
This creates a new arena for competition. Retailers are no longer vying solely for visibility on Google or online marketplaces; they are competing to be favorably interpreted by answer-generating machines. Messy product tagging and poorly written descriptions create gaps that AI models will fill with approximations drawn from external sources beyond a brand’s control. Given that LLMs can sometimes generate imperfect answers, brands should ensure they’re providing clear direction.
As such, more retailers are revisiting their product data taxonomies to improve consistency and ensure their attributes are intelligible to both humans and AI. If a retailer’s data does not provide a coherent story, AI systems won’t have much to work with.
Agentic commerce could accelerate these shifts. Open standards such as OpenAI and Stripe’s Agentic Commerce Protocol (ACP) provide a common language for AI agents to interact directly with ecommerce systems. Touted features include adding items to carts, verifying stock availability and applying loyalty points. As this space evolves, retailers that explore agent-to-site or agent-to-agent tools may uncover entirely new routes to reach customers.
Diligent Catalog Housekeeping Beats Trying to Game the System
The SEO model for product visibility has been a nonstop game of cat and mouse between retailers and ever-shifting search algorithms. Generative engine optimization (GEO) follows a similar pattern, with its proponents suggesting it can guide AI systems much like SEO once influenced search rankings. But with LLMs being even more of a black box than traditional ranking criteria, the fundamentals of GEO remain largely untested and untestable, with any findings always a step behind models that constantly update as they “learn.”
What is clear, however, is that retailers benefit from strengthening the same foundations that underpin product discovery in any channel. Structured data helps systems explain products, consistent feature labels support meaningful comparisons, and up-to-date pricing, stock and reviews enable recommendations. Retailers that invest in this much-needed catalog hygiene can enhance the shopping experience whether or not AI comes into play.
The risk that retailers face with end-to-end AI shopping is the same that open web publishers are currently experiencing: reduced referral traffic means being disintermediated from audiences, losing not only eyeballs and traffic but also the first-party data such visitors generate. This makes it pivotal to find ways to foster direct consumer interactions and source independent audience intelligence, such as insight derived from zero-party data, that is not beholden to any platform.
Query-based shopping ultimately relies on understanding customer motivations, which are now directly fed into the shopping journey. A deeper knowledge of current and prospective customers helps retailers present their products in ways that align with the factors that might motivate a purchase. Some brands and merchants are already incorporating more lifestyle-based product descriptions or proactively providing answers to anticipated queries directly within copy, rather than relying on filters alone.
The pace of change in commerce may feel relentless, but the path forward is clear and retailers unprepared for its direction risk being left behind. Observe, test, learn and forge bonds with your audiences that can endure any disruption.
With close to 20 years of experience in leadership roles, Anthony Flaccavento started his career working for high-profile consumer media brands, where he led all integrated sales efforts and was responsible for generating long-lasting relationships with major brands like Samsung, Verizon, Diageo, and Unilever. In his current role at Ogury, General Manager, Americas, Flaccavento is tasked with expanding Ogury’s business across the US, establishing the company as the go-to partner for all brands, retailers and marketers, helping them balance privacy with sustainable advertising results.
