The New Shopping Funnel Starts With AI
Something fundamental shifted in consumer behavior over the past 18 months. Instead of opening Google and typing "best running shoes for flat feet," a growing number of shoppers are opening ChatGPT, Perplexity, or Google's AI Overview and asking the same question conversationally.
The data backs this up. Similarweb reported that AI-referred traffic to retail and ecommerce sites grew over 700% between early 2024 and late 2025. Adobe Analytics found that during the 2025 holiday season, AI-referred visits to ecommerce sites generated 8% higher conversion rates than traditional search traffic. These visitors arrive with higher purchase intent because the AI has already pre-qualified the product for them.
But here's the problem for most Shopify merchants: AI shopping assistants don't recommend products the same way Google ranks pages. The signals that make you visible to AI are different from the signals that drive SEO. And right now, most stores are failing at the basics.
How AI Shopping Agents Decide What to Recommend
When someone asks ChatGPT or Perplexity for a product recommendation, the AI does not simply search the web and return the top Google result. The process is fundamentally different.
Training Data and Knowledge
Large language models were trained on massive amounts of web text, including product reviews, comparison articles, forum discussions, and editorial recommendations. Products that appear frequently in trusted sources -- Wirecutter, niche review blogs, Reddit threads, YouTube transcripts -- have a stronger presence in the model's knowledge. This means editorial coverage and user-generated content matter even more than they did in SEO.
Real-Time Search (Browsing Mode)
When ChatGPT browses the web (or Perplexity runs its search), it pulls information from live web pages. This is where your product pages, descriptions, and structured data become critical. The AI needs to be able to understand what your product is, who it's for, what it costs, and why it's worth recommending. If your product page is a wall of branded jargon with no clear specifications, the AI will skip it in favor of a competitor whose page is clearly structured.
Structured Data and Schema Markup
Schema.org Product markup is the machine-readable version of your product page. It tells AI systems (and search engines) the product name, price, availability, rating, review count, brand, and category in a standardized format. Without it, AI has to parse your HTML and guess. With it, AI gets clean data instantly.
According to an analysis by Ahrefs in 2025, approximately 48% of Shopify stores have either no schema markup or broken/incomplete implementations. These stores are functionally invisible to AI product searches that rely on structured data.
The Three Reasons AI Can't Find Your Products
1. Your Product Titles Are Brand-First, Not Search-First
This is the single most common mistake. A study of 10,000 Shopify stores by Profitero found that 41% use product titles that lead with internal branding, collection names, or creative labels that no customer would ever search for.
"Luna Glow Collection - Serenity"
"Hydrating Face Moisturizer with Hyaluronic Acid - 50ml"
"The Explorer - Carbon Black"
"Lightweight Waterproof Hiking Backpack 35L - Black"
When someone asks ChatGPT "what's a good lightweight hiking backpack," the AI is looking for pages that contain those exact terms in prominent positions: the title, the H1, the first paragraph. "The Explorer" tells the AI nothing. "Lightweight Waterproof Hiking Backpack" tells it everything.
This doesn't mean you can't have creative product names. It means your title should lead with the descriptive, searchable terms and include your brand name after. Your brand name is meaningful to returning customers but meaningless to AI doing product discovery for new ones.
2. Your Product Descriptions Don't Answer Questions
AI shopping queries are conversational: "What's the best moisturizer for dry skin in winter?" "What running shoes are good for plantar fasciitis?" The AI is looking for pages that directly answer these questions.
Most product descriptions focus on features and emotional language: "Crafted with care, our premium blend..." They don't state clearly: "This moisturizer is formulated for dry and sensitive skin. It contains 2% hyaluronic acid and ceramides for 24-hour hydration. Dermatologist-tested. Fragrance-free."
The second version answers the shopper's question. The first version sounds nice but gives the AI nothing to work with.
Practical tip: Add a "Who is this for?" section to every product description. State the use case, skin type, body type, experience level, or scenario explicitly. This is exactly the kind of language AI picks up on when matching products to queries.
3. You Have No (or Broken) Structured Data
Shopify themes include basic schema markup by default, but it's often incomplete or outdated. Many themes only include the product name and price, missing critical fields like:
- aggregateRating and reviewCount: AI heavily weights products with visible social proof. If your schema doesn't include your 4.7-star average from 238 reviews, the AI doesn't know about it.
- brand: Tells AI who makes the product. Important for brand-aware queries.
- material, color, size: These attributes are critical for filtering and comparison queries ("best cotton t-shirt" or "size-inclusive activewear").
- availability: AI won't recommend out-of-stock products if the schema says
OutOfStock. But if the schema doesn't report availability at all, AI might recommend the product and send the customer to a dead end.
To check your current schema, paste any product URL into Google's Rich Results Test (search.google.com/test/rich-results). It will show you exactly what structured data Google (and by extension, AI systems) can read from your page.
Five Things You Can Fix This Week
1. Rewrite Your Top 20 Product Titles
Follow this formula: [Product Type] + [Key Feature/Material] + [Size/Quantity] - [Brand Name]. Your top 20 products by revenue probably account for 60-80% of your sales. Start there. This alone will make your most important products visible to AI search.
2. Add a FAQ Section to Product Pages
Add 3-5 genuine questions and answers below each product description. Use the actual questions customers ask in emails, reviews, and chat. Mark these up with FAQ schema (FAQPage type). AI systems love FAQ content because it maps directly to conversational queries.
3. Fix Your Schema Markup
If your theme's default schema is incomplete, use a Shopify app like JSON-LD for SEO or Smart SEO to generate complete Product schema. At minimum, ensure your schema includes: name, description, image, price, currency, availability, brand, and aggregateRating (if you have reviews).
4. Don't Block AI Crawlers
Check your robots.txt file (yourstore.com/robots.txt). Some security apps and anti-bot tools block AI crawlers like GPTBot (OpenAI) and ClaudeBot (Anthropic). If you're blocking these, you're invisible to AI search. You want to allow:
User-agent: GPTBot Allow: / User-agent: ClaudeBot Allow: / User-agent: PerplexityBot Allow: /
Shopify's default robots.txt does not block these crawlers, but if you've installed a security or SEO app that modified it, check carefully. Some merchants have unknowingly blocked AI crawlers while trying to stop scrapers.
5. Build External AI-Readable Presence
AI models weigh information from multiple sources. If your product only exists on your Shopify store and nowhere else, it has a weak signal. Amplify it by:
- Getting listed on comparison and review sites in your niche
- Encouraging detailed customer reviews (on your site and on Google)
- Creating genuinely useful content on your blog that answers the questions shoppers ask AI (this article is an example of that strategy)
- Submitting your product feed to Google Merchant Center, which AI systems often reference
Monitoring Your AI Visibility
The tricky part is knowing whether any of this is working. Unlike Google where you can track rankings in Search Console, there is no "ChatGPT Console" that tells you when your product was recommended.
You can look for indirect signals: check your Shopify analytics for referral traffic from chat.openai.com, perplexity.ai, and other AI domains. Track UTM parameters that indicate AI-referred visits. Monitor your schema health regularly.
But this is manual, intermittent, and incomplete. You can see that someone arrived from ChatGPT, but you can't see the queries where your competitor was recommended instead of you. You can't see how AI systems describe your products versus alternatives. You're looking at your own data and missing the competitive picture entirely.
This is exactly the gap Glancefy Visibility is designed to fill. It periodically queries AI shopping assistants for product terms relevant to your catalog and reports back: which of your products are being recommended, which competitors show up instead, and what your product pages are missing (schema gaps, title issues, missing attributes) that prevent AI from surfacing them.
It also audits your structured data across all product pages, flags products with AI-unfriendly titles, and tracks your AI referral traffic over time so you can measure the impact of the optimizations you make.
The opportunity window: Right now, most Shopify merchants are not optimizing for AI discovery. The merchants who fix their titles, schema, and content structure today will capture a disproportionate share of AI referral traffic as it continues to grow. This window won't stay open forever. Once AI shopping optimization becomes mainstream, the advantage goes to whoever started earliest.
This Is Not a Future Problem
It's tempting to file "AI shopping optimization" under "things to worry about later." But the 7x growth in AI referral traffic happened in under two years. OpenAI launched ChatGPT shopping features in early 2025. Perplexity launched shopping in late 2024. Google's AI Overviews now surface product recommendations for commercial queries.
The shift from search-bar shopping to conversational AI shopping is happening now, not next year. And unlike SEO, where you're competing against every page on the internet, AI optimization is still early enough that basic fixes yield outsized returns. Most of your competitors haven't done anything. A week of focused work on titles, schema, and descriptions puts you ahead of the majority.
The question is not whether AI will influence how your products get discovered. It already does. The question is whether the AI knows your products exist.