Episode 05 — AI chat search is the new storefront

HOST A:  I want to start this episode with a thought experiment. When was the last time you actually typed a product question into a search engine, instead of just asking ChatGPT or Gemini directly?

HOST B:  Honestly… I can’t remember. It’s become the default for me without me even consciously deciding that.

HOST A:  That shift, happening quietly and at scale, is exactly what this episode is about. Today, nearly one billion prompts are issued daily to AI chat assistants.

HOST B:  A billion. Per day.

HOST A:  Per day. And here’s the number that matters even more for anyone in eCommerce specifically: 1.3 billion monthly referrals are flowing from those AI assistants to actual websites.

HOST B:  So this isn’t just people chatting with an AI in a closed loop. It’s actively sending traffic — and presumably purchase intent — out to retailers.

HOST A:  That’s exactly right. And the report frames this really well: for many shoppers now, discovery doesn’t begin in a search bar. It begins with a question to ChatGPT, or Gemini, or Amazon’s Rufus.

HOST B:  Which is a genuinely different starting point for the customer journey than what most retailers have built their entire digital strategy around for the past two decades.

HOST A:  Two decades of SEO thinking, built around the assumption that a human is typing keywords into Google and scanning a page of blue links. That assumption is breaking down, fast.

HOST B:  Let’s talk about why accessibility specifically is the hinge point here, because I think this is the least intuitive connection in the whole season for people who haven’t thought about it before.

HOST A:  So here’s the mechanism. An AI assistant — whether that’s a general one like ChatGPT, or a shopping-specific one like Rufus — needs to actually understand your product in order to recommend it, describe it accurately, or help a user complete a purchase through it. And the way it understands your product is by reading the underlying structure of your content.

HOST B:  Structured markup, alt text, accessible FAQs — the same kinds of things we’d talk about in the context of a screen reader user.

HOST A:  Exactly the same things. The report puts it directly: structured markup, alt text, and accessible FAQs allow products to surface in conversational recommendations. Without them, your catalog risks becoming invisible to both people and machines.

HOST B:  That phrase — invisible to both people and machines — feels like the thesis of this entire episode.

HOST A:  It really is. Because historically, if you neglected accessibility, the population you were failing was clearly defined: people with disabilities using assistive technology. Real, and a meaningful portion of your customer base, but a population some companies wrongly treated as a niche they could deprioritize. AI chat search changes that calculus completely, because now the exact same structural gaps that fail a screen reader user also fail the AI assistant that an enormous and rapidly growing share of all shoppers are starting their journey with.

HOST B:  So poor accessibility isn’t just an exclusion problem anymore. It’s becoming a visibility problem for your entire customer base.

HOST A:  That’s the shift. Let’s go through the benefits side, because the report lays out four specific categories, and I think they’re worth taking one at a time. First: visibility and SEO. Structured markup ensures AI assistants can actually cite your products.

HOST B:  Which is the new version of ranking on page one of Google, essentially.

HOST A:  Arguably more important, because if an AI assistant simply answers the user’s question directly with a recommendation, the user may never even click through to a traditional search results page at all.

HOST B:  What’s the second one?

HOST A:  Conversion. Reliable, accessible data improves trust and purchase follow-through. If an AI assistant can confidently and accurately describe your product because your data is clean and well-structured, that translates into a more confident recommendation, which translates into a shopper who’s more likely to actually complete the purchase.

HOST B:  Third?

HOST A:  Brand equity. Inclusive content strengthens your reputation across audiences — not just with the disability community, but broadly, because accessible content tends to simply be clearer, better-organized content.

HOST B:  And fourth, I’d guess, ties back to the last two episodes.

HOST A:  You’d guess right — compliance. Reducing exposure to lawsuits and regulatory penalties, which we already covered in depth in episode two.

HOST B:  So accessible, AI-ready content is solving four problems simultaneously with effectively the same underlying work.

HOST A:  Which is, again, the thread running through this entire season. Let’s flip it and go through the risk side too, because I think the downside framing lands just as hard. Invisibility — if your content isn’t structured properly, your products simply won’t appear in AI chat results at all. Not ranked lower. Not deprioritized. Absent.

HOST B:  That’s a stark way to put it.

HOST A:  It is, but I think it’s accurate, especially as more shoppers start their journey with an AI assistant instead of a traditional search. Second risk: exclusion. Shoppers with disabilities face barriers in discovery or checkout — the population this has always been about, still very much present and still very much affected.

HOST B:  And third?

HOST A:  Reputation loss. Failing to surface in this new AI-first marketplace signals irrelevance. If a competitor’s products are getting recommended by ChatGPT and yours aren’t, that’s not a neutral outcome — that’s actively losing ground to companies whose catalogs are simply more legible to these systems.

HOST B:  Let’s get into how eCommerce brands are actually responding right now, because I think the adoption numbers here are some of the most telling in the whole report.

HOST A:  They really are. Seventy-four percent of eCommerce businesses are already adjusting their product content specifically to be AI chat-friendly.

HOST B:  Three out of four. That’s not an early, speculative trend — that’s a majority behavior already underway.

HOST A:  And it goes further — another twenty-four percent are planning to make those adjustments in the near future. So you’re looking at roughly ninety-eight percent of eCommerce businesses surveyed who are either already adapting or actively planning to.

HOST B:  Which leaves almost nobody in the “not planning to do anything about this” category.

HOST A:  Functionally nobody — the report puts that figure at just one percent.

HOST B:  So if you’re listening to this and your company hasn’t started thinking about AI chat readiness yet, you’re not early to a trend. You’re meaningfully behind the vast majority of your competitors who’ve already started moving.

HOST A:  That’s exactly the right way to frame it. This has moved past the “should we even consider this” stage. The practical question for most companies now is “how fast can we close the gap,” not “should we get in line.”

HOST B:  Let’s bring this back to something concrete for anyone listening who wants a starting point. If structured markup, alt text, and accessible FAQs are the foundation here — where does a team actually begin?

HOST A:  I’d start with product pages, since that’s where AI assistants are most likely pulling information from when they’re making a recommendation. Make sure every product image has accurate, descriptive alt text — not just for screen readers, but because that’s literally the data an AI system uses to understand what the image is showing. Make sure product descriptions are written in clean, well-structured text, not buried inside an image or a PDF that a system can’t parse at all. Build out accessible FAQ content that directly answers the kinds of questions a shopper might ask an AI assistant. And make sure your underlying markup — the structured data behind the page — accurately reflects price, availability, and specifications, because that’s frequently the layer AI systems rely on most heavily.

HOST B:  And all of that, notably, is also exactly what you’d do to make that same page work well for a screen reader user.

HOST A:  Which is, I think, the cleanest possible illustration of this entire season’s argument. We’ve now gone through lawsuits, cart abandonment, voice commerce, and AI chat search — four completely different business problems, sitting in four different departments, on four different parts of the P&L.

HOST B:  Legal risk. Conversion and revenue. An entirely new growth channel. And now, fundamental visibility in an AI-first marketplace.

HOST A:  And in every single case, the underlying fix is the same work: building genuinely accessible digital experiences.

HOST B:  That’s a strong place to pause for this episode. Next time we’re heading into regulation again, but at a different altitude — the European Accessibility Act and UK requirements, and what it means that compliance is now going global, not staying confined to any one country’s legal system.

HOST A:  As of June 2025, the EAA is already in effect, and seventy-one percent of eCommerce leaders surveyed say it’s already affecting how they operate. We’ll dig into what that actually requires, and why this matters even if you’re not based in Europe.

HOST B:  See everyone there.

HOST A:  Thanks for listening.