We’re starting to see a shift in how traffic is discovered, especially with AI-powered search like Google SGE and tools like ChatGPT summarising results instead of sending users directly to websites.
For Shopify store owners, this raises a few questions:
Are product pages still getting the same organic visibility?
Is traditional SEO enough anymore?
How do we make sure our content/products are actually surfaced in AI-generated answers?
From what I’m seeing, structured content, entity clarity, and topical authority seem to matter more than ever — but it’s still pretty unclear what’s actually working in practice.
Curious to hear from others:
Have you noticed any drop or shift in organic traffic recently?
Are you adapting your SEO/content strategy for AI search?
Has anyone tested optimising for AI visibility specifically?
Feels like this could be a pretty big shift for eCommerce if it keeps going this way.
I’ve been watching this closely and yeah, it’s real. The click-through from Google has been dropping for a while now, even before SGE. Featured snippets already took a chunk, and AI overviews are accelerating that.
What I’ve noticed so far with the stores I work with: informational queries got hit hardest. Stuff like “best [product] for [use case]” used to drive a lot of traffic. Now Google just answers it directly. Product-specific and brand queries are holding up better because people still want to see the actual product page.
A few things that seem to help. Making sure your product data is structured properly (JSON-LD schema, clean titles, real specs instead of marketing fluff). AI models tend to pull from well-structured sources. Also having actual unique content on product pages rather than manufacturer descriptions that 50 other stores also use.
The biggest shift I think is that “being findable” isn’t just about ranking on Google anymore. You need to show up wherever people are researching. That could be Reddit, YouTube, TikTok, or even ChatGPT conversations. Diversifying traffic sources feels less optional now.
Have you noticed the drop more on specific product categories or across the board?
Yeah this is something I’ve been looking into quite closely actually. The short answer is yes — and the data backs it up pretty clearly.
Shopify reported in their Q4 2025 earnings that merchants with comprehensive Product schema markup saw 34% higher inclusion in AI shopping features. The problem is only about 12% of Shopify stores actually have it set up properly.
For supplement and health brands its even more of a challenge because AI treats these as credibility-sensitive categories. It wants structured data, third-party citations, and evidence-based content before it’ll recommend you. So the bar is higher than most product verticals.
The fix isn’t more SEO content — it’s restructuring your product data so AI can actually parse it. JSON-LD schema (Product, FAQ, Review) is the single fastest techincal win. Then building what I’d call “discovery pages” that answer the actual questions people ask AI — things like “best magnesium for sleep” rather than just “magnesium supplement.”
Happy to share more on what’s working if anyone’s intrested. We’ve been testing this across a few hundred queries and the patterns are pretty consistent.
These are the right questions. I’ve been testing AI visibility specifically for the last few months and here’s what I’m finding:
Traditional SEO isn’t enough anymore, but it’s still the foundation. Everything that makes your store rank well on Google clean site structure, proper meta tags, fast loading, good content also helps AI agents understand your store. The difference is that AI agents go a step further. They don’t just index your pages they try to understand what you sell, who it’s for, and whether to recommend you over competitors. That requires a different layer of optimization.
What’s actually working in practice for AI visibility:
Structured data matters more than ever. JSON-LD schema markup on your product pages tells AI exactly what your product is price, availability, reviews, attributes. Most Shopify stores have basic schema from their theme but it’s often incomplete or missing key attributes. Check your pages with Google’s Rich Results Test if your structured data is thin, AI agents are getting a thin picture of your products too.
Product descriptions need to be attribute-rich, not just persuasive. AI agents comparing products across stores are looking for specific details materials, dimensions, ingredients, compatibility, use cases. “Beautiful handcrafted mug” tells AI nothing useful. “12oz ceramic mug, hand-thrown stoneware, food-safe glaze, dishwasher safe, made in Vermont” gives AI everything it needs to match your product to a shopper’s query.
llms.txt is a newer signal worth paying attention to. It’s a file you host on your domain that gives AI agents a structured overview of your store think of it as a resume for your business that AI reads before deciding whether to recommend you. It’s still early but stores with llms.txt are giving AI agents more context than stores without one.
On the traffic shift question the honest answer is that most Shopify merchants can’t measure AI-referred traffic yet because it doesn’t show up as a distinct source in analytics. It typically appears as direct traffic or referral from chatgpt.com. So the drop you might be seeing in organic could partially be traffic shifting to AI channels without being properly attributed.
A practical test anyone can do right now: Go to ChatGPT and ask “recommend me a store for [your product category].” If your store shows up, look at how it describes you is it accurate? If it doesn’t show up, that’s your starting point for optimization. Then do the same on Perplexity and Gemini. The results vary across platforms, which tells you a lot about which AI agents have better access to your data.
Would be interested to hear if anyone else has done this test and what they found.
Yes, but the way most people frame this is slightly off — and it’s worth being precise about. Shopify activated Agentic Storefronts for every eligible merchant on 24 March, so most stores are technically already inside ChatGPT. The visibility problem isn’t enrolment, it’s that the feed is a list, not an answer.
ChatGPT builds recommendations from editorial roundups, review sites, Reddit and Bing’s top organic. About 87% of ChatGPT Search citations match Bing’s top results, so if your brand isn’t in those third-party sources, the feed alone won’t carry you.
Three things worth checking today:
robots.txt — a fair few Shopify themes block GPTBot, ClaudeBot and PerplexityBot by default. If you’re blocked there, nothing else matters.
Product schema completeness. Not just the basics — FAQ, glossary entities, ingredient-level markup if relevant.
Whether anyone outside your domain says you’re the answer to the question buyers actually type. If no third party corroborates it, the model has nothing to anchor on.
Yes, we have noticed a discovery shift from clicks to answers. Product pages continue to rank, but thin manufacturer copy and weak collections will lose visibility fast. Old-school SEO still works, but it has to be supplemented by other methods. For Shopify store: focus on product schema, FAQs, comparison content, shopping guides, clear brand entities, unique product description and so on. Create topic clusters for categories so your site can become the #1 reference. You should also monitor changes in impressions in your Search Console even if clicks have dropped. I would also treat the collection pages as landing pages and optimize these too along with your product pages. AI tools prefer clear, authoritative, well-structured information that can be distilled and cited back to a single source in a uniform manner.
I have been watching the same shift. Traditional SEO is not dead, but it is not enough anymore. AI models like ChatGPT and Google SGE pull answers from sources they trust, not just the highest ranking pages.
What has worked for me is focusing on entity clarity and structured content. You need to tell search engines and AI what your product is, who it is for, and why it is the best answer. Not just keywords.
A friend in ecommerce started working with Fortress on generative engine optimization. They helped him restructure his product pages to be more AI friendly. His traffic from ChatGPT referrals went up, not down.
The shift is real. But stores that adapt early are the ones who will win. If you are on Shopify, start by cleaning up your product schema and writing content that answers specific questions, not just describes features.
It feels like generic product pages are becoming much weaker in AI driven discovery.
Stores using the same manufacturer descriptions, thin product details, or vague copy all start blending together once AI generated answers become part of the shopping flow.
The stores that seem strongest now are usually the ones adding real context, clearer explanations, comparisons, FAQs, or actual buying guidance around the products.
Great questions, and the honest answer to “is traditional SEO enough” is no, but usually not for the reason most posts give.
Getting recommended by AI search runs on two separate systems, and most stores only fix one. System one is your schema and product data, what the AI reads about each product. System two is crawler access, whether the AI’s crawler is even allowed to fetch your pages in the first place. You can have flawless schema and still be invisible if the crawler is blocked. This matters more now that stores are getting auto-connected to ChatGPT shopping, because being connected is not the same as being readable.
The part that trips people up: the crawler that matters for ChatGPT is not GPTBot. GPTBot is OpenAI’s training crawler. The one that decides whether you show up in ChatGPT’s search results is OAI-SearchBot. Most guides say “allow GPTBot” and stop, which is the wrong lever. Same pattern elsewhere: PerplexityBot for Perplexity, and Google-Extended only affects Gemini, not Google’s AI Overviews.
Two free checks, no tools needed:
Open yourstore.com/robots.txt and look for any Disallow rules hitting OAI-SearchBot, PerplexityBot, or Google-Extended. If those are blocked, fix that first.
On a product page, right-click, View Page Source, and search the raw source for application/ld+json. If your schema shows up in the live page but not in the page source, JavaScript is injecting it, and a lot of AI crawlers will not run that JavaScript, so they never see it.
On your point about structured content and topical authority, both genuinely help, but they only pay off if the crawler can reach the page and read the schema in the raw HTML first. Fix access and readability, then content.
Full disclosure: I build a Shopify app in this space, so this is what I stare at all day. Happy to dig into anything specific about your store’s setup.
Lumine’s point about diversifying traffic sources is the right framing. The stores adapting fastest aren’t treating AI visibility as an SEO extension, they’re treating it as a separate channel with its own signals, its own testing methodology, and its own version of “ranking.”
The thing that catches most merchants off guard: your Google rankings and your AI visibility can point in completely different directions. A store ranking well on Google for a category term might be invisible in ChatGPT or Perplexity for the same query. And vice versa - there are stores with fairly modest SEO footprints showing up consistently in AI answers because they happen to be mentioned in the sources those models weight heavily (Reddit threads, editorial roundups, niche forums).
The drop in click-through from informational queries you mentioned is real, and it’s probably permanent for that query type. The stores that will weather it best are the ones building presence in the places AI pulls from. Which, right now, means third-party mentions on sources the models trust, not just better on-site content.
The diagnostic most merchants haven’t done: run your 5 most valuable traffic-driving search queries through ChatGPT and Perplexity as if you were a buyer. See who appears. If it’s consistently the same 2-3 competitors, look at where they’re mentioned that you’re not. That gap is more actionable than another round of on-site fixes.