AI Is Eating Search — Email and Structured Data Are Your Moat

Most of what you have read about AI search this quarter is half right. Two shifts are running in parallel. AI-mediated search (Google AI Overviews, ChatGPT shopping, Perplexity) is answering an increasing share of commercial queries before the user ever reaches your site. At the same time, the acquisition channels brands relied on for the last decade (organic search, social algorithms, paid discovery) are degrading. SparkToro's data puts zero-click searches at roughly 55% in 2025. The brands that hold position over the next eighteen months will probably not be the ones with the biggest ad budgets. They will be the ones who built two specific assets ahead of the curve. An owned audience through email. Product data structured well enough that AI systems can actually consume it. The arbitrage is real, though smaller than the loudest pundits claim. Cleaner than most arbitrages in commerce, though.

Who Loses: Brands Still Playing the Old Discovery Game

SparkToro's read on the zero-click era is blunt. No amount of tactical excellence in content marketing, social posting, or keyword optimization saves you when the platform keeps the click for itself. Google summarizes your blog post in the AI Overview. Instagram buries your product carousel under whatever the For You algorithm wants today. TikTok moves on. If your acquisition strategy depends on rented audiences and algorithmic grace, you are running a business on borrowed time. Practical Ecommerce's analysis of product detail pages catches the other half of the problem. Most brands still treat the PDP as a static brochure. Plenty of lifestyle copy. Almost no structured, entity-level data. The result is a quiet exclusion. AI shopping agents skip pages they cannot parse. Competitors who format size, material, use case, and compatibility as structured data get surfaced. You do not. The losers in this window are not bad brands. They are brands with infrastructure that has not aged well.

Who Wins: Operators Building Owned Channels and AI-Ready Product Intelligence

Email turned 55 this year. The channel has never been more strategically useful. When every discovery algorithm is a black box that changes quarterly, email is the one direct line you control end to end. No intermediary decides whether your message reaches your customer. The D2C brands that survived the pandemic boom-and-bust cycle did it the same way. They converted early hype into durable subscriber relationships. They stopped chasing viral moments and started compounding a list. On the product data side, the operators ahead of the curve are treating every PDP as a structured knowledge node. Schema markup. Attribute-level detail. Machine-readable specifications. The point is to let AI shopping tools parse, compare, and recommend with high confidence. This is not busywork. It is the new SEO. Except instead of optimizing for a results page with ten blue links, you are optimizing for an answer engine that may never show a link at all.

Three Questions to Pressure-Test Your Moat

Three questions a CTO and a CMO should be able to answer together before approving the next quarter's discovery budget. One. If I pull your top 20 PDPs right now and run them through Google's Rich Results Test, how many pass cleanly with structured schema for price, availability, material, size, and use case? If most fail, your AI shopping visibility is roughly proportional to luck. Two. What is the conversion rate on your owned email list compared to your blended paid acquisition cost over the last 90 days? If email lags badly, you have probably underinvested in the only channel that does not depend on a platform's goodwill. Three. When you search your own top 10 SKUs inside ChatGPT, Perplexity, and Google AI Overviews today, where do you appear, where do you not, and what is the structural reason for each absence? Without that map, your roadmap is based on intuition. What would change my read on this is a brand that locked in durable AI search visibility without structured data investment. I have not seen one. Until then, the calibrated bet is structured data plus owned audience, in that order.