The world of e-commerce is standing on the precipice of its most significant shift since the invention of the online shopping cart. We are moving from an era of static search and scrolling to an era of “Agentic Commerce”—a world where Artificial Intelligence agents not only assist in discovery but actively curate, reason, and guide purchasing decisions.
In a recent deep-dive discussion, Yotpo CEO Tomer Tagrin and VP of Strategy Itai teamed up with Shopify’s builders—Product Director Ellen Dunne and VP of Product/AI Mujtaba Khambatti. Together, they explored how AI is fundamentally altering the landscape for shoppers and brands alike. From the death of traditional keywords to the rise of conversational storefronts, the insights shared offer a roadmap for merchants trying to navigate this brave new world.
If you are a merchant, a marketer, or a developer, this is not a trend you can afford to ignore. As the panel noted, even if only 10-15% of commerce moves to agentic flows in the next three years, that represents a massive shift in consumer behavior—equivalent to retail opening up on the moon.
The first major topic addressed by the panel was the evolution of product discovery. For the last two decades, discovery has been dominated by the search bar. You type in a keyword, you get a list of links, and the burden of research falls on you. You open ten tabs, compare specs, read reviews, and hope for the best.
That behavior is changing rapidly. As highlighted in the discussion, nearly 40% of U.S. consumers are already using generative AI for discovery. The difference lies in the fidelity of the interaction. In the old world, you searched for “best running shoes.” In the new world of Agentic Commerce, a user says, “I am training for a marathon, I have flat feet, and I run mostly on pavement. What should I buy?”
Mujtaba Khambatti, VP of Product/AI at Shopify, noted that traffic coming from these AI interactions is fundamentally different. When a user clicks through from an AI agent to a merchant’s storefront, they are often ready to buy. Why? Because the “consideration phase”—the messy middle of the funnel where customers usually drop off—has been handled by the agent.
For brands, this means that showing up in these AI conversations is paramount. If an AI is weighing options for a luxury watch or a vegan leather bag, your brand needs to be part of that consideration set. If the AI doesn’t understand your brand’s context—that you are sustainable, or luxury, or specifically designed for flat feet—you simply won’t exist in the results.
Ellen Dunne, Product Director at Shopify, emphasized that this shifts the game from keyword stuffing to context building. It’s no longer about ranking for a generic term; it’s about answering complex, natural language questions. The AI is acting as a disambiguation engine, filtering through the noise to find the perfect match. Brands that can provide rich, descriptive context about their products are the ones that will win in this new environment.
If discovery happens off-site in an LLM, what is the role of the Product Detail Page? The panel suggests that the PDP is undergoing a radical transformation. Historically, PDPs were designed exclusively for human cognition. Designers obsessed over “above the fold” content—making sure the image, price, and buy button were immediately visible to reduce friction.
In an agentic world, the PDP serves a dual purpose. It must still be beautiful and conversion-optimized for the human who lands there, but it must also be a data-rich repository for the AI agent that crawls it.
Ellen Dunne drew a fascinating parallel to old-school web design. While visual clutter is bad for humans, AI agents have a voracious appetite for text and detail. This gives merchants the freedom to place extensive information “below the fold.”
Details that might bore a casual browser—manufacturing processes, supply chain ethics, detailed material breakdowns, or extensive FAQs—are gold for an AI. When an agent scans your site, it looks for these specific details to answer user queries accurately. If you hide this information or leave it out for the sake of minimalism, you are effectively silencing your brand in the AI conversation.
To help merchants bridge this gap, Shopify introduced tools like the Knowledge Base app. This allows merchants to view and edit the “facts” that Shopify knows about their store. It’s a way to establish a “Ground Truth.”
Mujtaba explained that AI can sometimes “hallucinate.” If a user asks about a return policy and the AI can’t find a specific answer on your site, it might guess based on industry standards (e.g., “It’s probably a 30-day return”). By explicitly structuring this data, merchants gain a measure of control over how their brand is represented, even on platforms they don’t own.
One of the most profound insights from the discussion was the changing nature of social proof. There is an old adage in e-commerce that no user reads review number 1,027. Humans look at the aggregate star rating, read the top three reviews, and maybe filter by “most recent.”
AI, however, reads everything. It is a “super-reader.”
Tomer Tagrin pointed out that Yotpo has invested heavily in extracting attributes from reviews for this very reason. A specific product might have thousands of reviews, but perhaps only three of them mention that the product is “great for psoriasis.” A human user would never find those three needles in the haystack.
However, if a user asks an AI agent, “Find me a lotion that works for psoriasis,” the agent can instantly recall those specific reviews and surface the product. This makes the quality and specificity of user-generated content (UGC) more valuable than ever. It’s not just about “great product” anymore; it’s about the context of the usage.
We live in an era where consumers are increasingly skeptical of what they see online. Is this photo real? Is this review a bot? Is this video a deepfake? In this environment, authenticity becomes the ultimate currency.
The panel discussed how AI agents triangulate data. They look at the merchant’s site, but they also cross-reference with Reddit, YouTube, and third-party blogs. If your customer service is terrible, an AI will know because it has read the complaints on Reddit. This means reputation management is no longer just about your own site; it’s about the holistic footprint of your brand across the web.
While off-site discovery is huge, the on-site experience is also becoming more conversational. Users are being trained by ChatGPT to expect answers to their specific questions. When they land on a static website that forces them to hunt for information, it feels archaic.
Shopify is rolling out Storefront MCP (Model Context Protocol) capabilities that allow merchants to embed conversational agents directly into their stores. Ellen Dunne described this as hiring a digital sales associate. Just as you would train a retail employee on your brand voice, tone, and product specifics, you can train your storefront agent.
This allows for a level of personalization previously impossible online. If a customer lands on a site and asks, “I’m looking for a gift for my wife who loves gardening but hates clutter,” a standard search bar fails. A storefront agent, however, can reason through the catalog and suggest the perfect, compact gardening tool set.
Mujtaba shared a compelling story about a merchant selling Indian wedding attire. Indian weddings are multi-day, complex, emotional events involving diverse outfits for different ceremonies. This merchant implemented a storefront agent to guide families through the process.
Instead of feeling overwhelmed by a catalog of thousands of items, customers felt like they were talking to a knowledgeable “auntie” who guided them on what to wear for the Haldi ceremony versus the reception. This didn’t just help navigation; it built an emotional connection. The website became more than a catalog; it became a consultant.
So, how does a brand actually prepare for this shift? The consensus from the panel was clear: urgency and experimentation are key. You cannot wait for a “best practices” playbook to be written because the technology is evolving too fast.
In a world of infinite AI-generated content, your unique brand story is your moat. Why did you start the business? Who makes the products? What are your values? As Ellen noted, these emotional hooks are what convert buyers once the AI has done the logical filtering. Ensure this story is written down, structured, and accessible on your site.
Itai, VP of Strategy at Yotpo, suggested viewing this era much like the early days of SEO. There is no perfect “Google Analytics” for LLMs yet. You have to experiment. Try different product descriptions. Add rich FAQs. Monitor your chat logs to see what people are actually asking, and then update your content to answer those questions proactively.
For small brands and entrepreneurs, this is a golden era. Trying to beat Nike on the keyword “running shoes” in Google is impossible. But winning on a specific, long-tail query inside a chat interface is entirely possible if your data is better and your product fit is more precise.
Use the conversational data you collect to improve your products. If users keep asking your AI agent if your backpack fits a 16-inch laptop, and you don’t have that data, go measure it and add it. If they ask for a color you don’t have, pass that to product development. The conversation is the highest fidelity data source you will ever have.
We are witnessing a reset of the commerce playground. The incumbents who dominated the last decade of SEO and paid ads do not automatically win this next round. The winners will be the brands that are most transparent, most authentic, and best structured for the AI age.
As Tomer Tagrin concluded, the most important strategy right now is to stay curious. Read newsletters (like Commerce GPT), experiment with Shopify’s new tools, and don’t be afraid to fail fast. The train is leaving the station, and it’s headed toward a world where commerce is conversational, contextual, and agentic. Whether you are a massive enterprise or a garage startup, now is the time to optimize your brand for the machine-driven future.
Agentic Commerce refers to a new phase of e-commerce where AI agents (like chatbots or LLMs) act on behalf of the user to research, compare, and discover products, rather than the user doing all the manual searching themselves.
Focus on creating rich, detailed content. Use the Shopify Knowledge Base app to ensure your store’s facts (FAQs, policies) are structured and accurate. Ensure your Product Detail Pages include extensive details “below the fold” regarding materials, usage cases, and brand story.
Not immediately, but it is shifting the landscape. While traditional SEO focuses on keywords and backlinks, optimization for AI (sometimes called GEO – Generative Engine Optimization) focuses on context, brand authority, and answering complex natural language queries.
AI models read and analyze all reviews to understand specific product attributes. Having authentic, detailed reviews helps your product surface for niche queries (e.g., “best moisturizer for dry skin”) that generic product descriptions might miss.
A storefront agent is an AI-powered chatbot embedded on your e-commerce site. Unlike old chatbots that used rigid decision trees, these agents use LLMs to have natural, fluid conversations with shoppers, guiding them to products and answering specific questions just like a sales associate.
Itai Bengal
Director of Product Partnerships @ Yotpo
Our latest news and trending topics
