AI visibility becomes the new battleground in consumer discovery

If AI cannot read your catalog, it cannot sell your brand. That is the real wake-up call in Lisa Curtis’s argument: the brands disappearing from ChatGPT, Gemini, Claude, and Perplexity are not necessarily weak brands, they are unreadable ones.

AI discovery is replacing the old page-one game

For years, consumer discovery was shaped by Google rankings and Amazon placement. OpenAI now describes ChatGPT as a place where people explore, compare, and discover products, with shopping research designed for decisions involving comparisons, trade-offs, or multiple constraints. That is a different job from classic search, and it changes what wins visibility.

The scale makes this impossible to ignore. OpenAI says more than 700 million people turn to ChatGPT each week for help with everyday tasks, including finding products they love. Once a tool with that kind of reach becomes a first stop for product research, AI visibility stops looking like a novelty and starts looking like a distribution channel.

Why brands go missing

The most common mistake is assuming AI systems retrieve brands the way search engines do. They do not. OpenAI’s commerce architecture says ChatGPT ingests structured catalog data and product feeds to surface relevant products in context, which means the model is looking for machine-readable signals, not just brand storytelling.

That is why a household name can still vanish in an AI answer. Curtis’s brand audit surfaced a surprising visibility result even with a broad catalog, and that is the uncomfortable part for a lot of teams: human awareness does not automatically translate into model legibility. If the product data is thin, inconsistent, or hard to map, the AI system has very little to work with.

The diagnosis is simple, even if the fix is not. Brands built for the old search world often optimize for impression, voice, and broad category language, while AI systems want structured attributes they can compare at speed. If your catalog does not clearly tell the machine what the product is, who it is for, and how it differs from alternatives, you are effectively handing the recommendation to someone else.

What the systems reward

The new gatekeepers reward structure. OpenAI says merchants provide a structured product feed so ChatGPT can accurately surface products in search and shopping experiences, and its March 24, 2026 update says ACP is being expanded to support product discovery with more complete, relevant, and up-to-date information. In plain English, the cleaner the feed, the easier it is for the model to place your product in the right context.

Google has been saying something similar for years, which is why this shift is not as mysterious as it looks. Google Search Central says product structured data can help product information appear in richer search experiences across Google Search, Google Images, and Google Lens. Perplexity adds another clue: its shopping results appear as product cards with key details and are not sponsored, a sign that AI shopping is moving toward structured, source-backed presentation instead of simple keyword slots.

What matters here is not just presence, but legibility. Clean identifiers, accurate pricing, inventory status, and attribute mapping are doing more work than glossy copy ever will. If the system cannot tell which version of a product is in stock, what problem it solves, or how it differs from the next option, it will default to something else.

The transactional turn is already here

This is not theoretical. OpenAI launched Instant Checkout in ChatGPT on September 29, 2025, starting with U.S. Etsy sellers and with over a million Shopify merchants coming soon. That move turned product discovery into a path toward purchase, which raises the stakes for every catalog field a merchant controls.

The March 24, 2026 expansion of ACP reinforces the same point. OpenAI is not just trying to help people browse; it is building the plumbing that lets better product data flow into the answer surface itself. Brands that treat AI discovery as a side project are going to keep losing ground to competitors that treat feeds, metadata, and catalog hygiene as core commerce infrastructure.

What to change first

Start with the catalog, not the campaign. The fastest wins usually come from auditing whether every product can be read cleanly by a machine: titles, descriptions, categories, variants, pricing, inventory, and the attributes shoppers actually use to compare options. If any of that is vague or inconsistent, fix it before you spend another dollar polishing brand copy.

Then make sure the signals outside your own site support the story you want AI to tell. That means structured product data on your pages, consistent product feeds, and enough third-party reinforcement that the model can trust what it sees. The goal is not to trick the system; it is to make your product easy to understand when the system is deciding what to recommend.

Curtis matters here because she is not speaking from a theory deck. As founder and CEO of Kuli Kuli Foods, a Certified B Corp selling superfood powders, gummies, and lattes in 11,000 stores, she lives in the kind of consumer commerce environment where discovery gets fragmented across channels. Her point is blunt because the market is becoming blunt: if AI cannot parse your offer, it will hand the customer to someone else’s.

The next fight in consumer discovery is not about louder branding. It is about whether your brand is visible at all when a model assembles the answer. In that world, structured data is no longer technical cleanup; it is shelf space.