Agentic Commerce: The Future of the Digital Shelf. How Shopping is shifting from browsing to delegation.

The talk argues that online shopping is moving from humans searching, browsing and comparing to delegating the task to AI agents. The agent researches, compares and eventually decides what to buy. The speaker, an e-commerce agency practitioner, says this reduces brand exposure because there are fewer places to insert ads or competing products, and that only one or a few brands will win a given AI query. His summary is that if you are not recommended, you do not exist. He claims 40–60% of visibility comes from three factors: how content is structured, how clear the information is, and how relevant the product is to the question. The source gives no evidence for this figure. A large part of the talk contrasts human and machine reading. People respond to emotion, visuals, packaging and brand stories (Tony's Chocolonely is the example), while AI agents look for structured, consistent, factual data. For instance, "contains 10,000 collagen peptides per serving" is easier to compare than "premium quality formula." The speaker says agents extract product type, attributes, use cases, price versus value, and review sentiment across retailers. He recommends bullet points containing facts, comparison tables and data sheets, and FAQs that mirror real questions. He advises keeping SEO and adding GEO (optimizing for AI answers), with clear benefit and attribute hierarchies and machine-readable formats. The speaker argues that AI already holds a perception of each brand, and that brands often don't know how it describes them, which competitors it compares them to, or when and why they are recommended. He presents his agency's tool, which takes brand and product data plus prompt intents and returns scores across LLMs, competitor insights and suggested actions. In one customer example, the brand was well recommended in ChatGPT but still had work to do in other AI tools. He describes the method as an iterative cycle of changing content, re-auditing and retesting, because the field changes often. The closing advice, usable without the tool, is a three-step framework. First, be discoverable by including the phrases people use and stating category, audience and use. Second, be comparable through clear specifications, standardized attributes consistent across your site, retailers and marketplaces, and side-by-side modules. One client's product was listed as machine washable on one site but not another. Third, be recommended through strong sentiment, social proof and unique benefits. The talk is largely practitioner opinion and experience, with no quantified results beyond the visibility figure.