The search result used to be the answer. You typed in what you wanted, Google showed you a page of results, and you clicked the most relevant one. That model dominated product discovery for two decades.
Something different is happening now. A growing number of shoppers are describing what they want to an AI - "I need a durable waterproof backpack for hiking under $200" - and getting back three specific product recommendations, with reasons. No ten blue links. No ad auction. Just a curated shortlist.
That's Shopping AI. And while it won't replace Google Shopping or paid search anytime soon, it's already changing how products get discovered - and which businesses benefit.
If you're running Google Shopping campaigns or Performance Max, you have more at stake here than most. Here's what's actually happening, and what to do about it.
What "Shopping AI" Actually Means
Shopping AI refers to the integration of AI-driven recommendation and purchase capabilities into platforms that consumers already use daily. Right now, there are three main surfaces you need to understand.
Google AI Mode + AI Overviews
Google's AI Overviews appear on 14% of shopping queries. AI Mode - Google's conversational search layer - had 75 million daily active users as of February 2026. Includes sponsored store placements alongside organic AI recommendations.
ChatGPT Shopping
Launched 24 November 2025. Surfaces product recommendations with images, pricing, and links. Shopify and Etsy merchants auto-integrated. Others can submit product feeds via chatgpt.com/merchants. Currently organic - no bid required.
Perplexity Shopping
"Buy with Pro" for US Perplexity Pro users. Powered by web crawl and structured data - no direct feed submission. The same product data hygiene that helps you in Google AI Overviews applies here too.
Each works differently, but they share a defining characteristic: they're recommendation engines that pull from structured product data, review signals, and web content. They don't show you an ad auction. They surface products they assess as the best answer to a specific need.
Why This Matters for Your Ad Account
Most advertisers are still operating as if product discovery happens exclusively through keyword auctions. The numbers say otherwise.
ChatGPT Shopping traffic converts at approximately 15.9% compared to 1.8% for Google organic. The reason: AI-filtered visitors arrive pre-qualified. They've described their need to an AI, had it process their requirements, and been pointed to specific products. By the time they reach your site, they've already done more pre-purchase work than most search visitors.
Volume is modest - ChatGPT drives under 1% of sessions for most retailers right now. But at that conversion rate, the revenue contribution punches above its traffic weight. And the trajectory is up.
For Google specifically, the impact is immediate. AI Overviews on 14% of shopping queries means a significant portion of your category's traffic now interacts with AI-generated content before seeing your ads. Google expanded AI Overviews to Australia in December 2025. If you're running Shopping campaigns here, this is already your reality.
Here's the critical shift: traditional Shopping visibility was almost entirely bid-driven. AI surface visibility adds a different variable - feed quality and product data completeness. An advertiser with a thin, incomplete feed will be deprioritised on AI surfaces regardless of budget. The actions that improve your AI surface visibility are the same actions that improve your Shopping campaign quality. They compound.
One thing worth flagging: Google's loyalty programme annotations went live in Australia in March 2026. If you have a loyalty or rewards programme, member pricing and exclusive delivery options can now surface directly in AI Mode and Gemini. Most advertisers haven't activated this yet.
How to Get Found on Each Platform
Google AI Mode and AI Overviews
Start with your Merchant Center feed. A clean, verified feed with no disapprovals is the baseline. Beyond that, these are the levers that move the needle on AI surfaces:
Product titles: Use the structure Brand + Product Type + Key Differentiator + Model/Size/Variant. "Patagonia Down Parka Hooded Waterproof 600-Fill" outperforms "Patagonia Jacket." AI systems match natural language queries to structured product attributes - descriptive titles bridge that gap.
GTINs and MPNs: Every product with a barcode needs its GTIN in your feed. Google uses these to verify product identity and match your listing against third-party review data. Missing GTINs are one of the most common reasons products are deprioritised in competitive categories.
Images: High-resolution, multiple angles. AI surfaces with visual components favour products with image quality that meets category standards.
Loyalty annotations: If you run a customer loyalty programme, activate it in Merchant Center. Member pricing and exclusive delivery options now surface in AI Mode and Gemini in Australia.
Universal Commerce Protocol: Google's open standard enabling AI agents to execute purchases across retail platforms. Shopify, Etsy, Wayfair, Target, and Walmart are already integrated. Shopify merchants are part of this ecosystem automatically. For other platforms, watch for integration announcements - this is where the infrastructure is heading. For more on how AI agents interact with product data, see how AI shopping agents discover products.
Diagnostics: Check Merchant Center weekly for disapprovals. Price mismatches, missing identifiers, and policy violations remove products from all surfaces - including AI. Fix them within 48 hours.
ChatGPT Shopping
Shopify and Etsy merchants are already connected automatically. No setup required. For every other platform:
- Apply at chatgpt.com/merchants to submit your product feed
- Format to OpenAI's feed spec - required fields:
feed_id,account_id,target_merchant,target_country, and per-productidandtitlewithin the variants array - Add recommended fields: product descriptions, pricing (
price,list_price), availability, variant options (colour, size), seller information, barcodes
ChatGPT's recommendations are influenced by review signals. A product with 200 verified reviews will outrank a technically similar product with 12 reviews. If you're not actively collecting reviews and getting them structured into your data, you're leaving visibility on the table.
Current visibility is organic - not auction-based. That's a genuine advantage for well-structured, well-reviewed products. It won't stay that way permanently as OpenAI scales its ad product, but right now the playing field is unusually level.
Perplexity Shopping
No direct feed submission. Perplexity crawls the web and surfaces products based on structured data, review mentions, and web authority signals. The practical checklist mirrors your general AI readiness work:
- Schema.org Product markup on every product page
- Accurate, current pricing and availability
- Descriptions that answer "is this right for [specific use case]?"
- Reviews structured with AggregateRating schema
Perplexity Shopping is currently US Perplexity Pro only, but expanding. Getting your data hygiene right now means you're positioned when it reaches your market. For a deeper look at catalogue structure for AI visibility, see structuring your catalogue for AI visibility.
What This Means for Your Google Ads Strategy
Shopping AI doesn't make Google Ads less important. It makes your Google Ads feed more important.
Google AI Mode already includes sponsored store placements - live as of February 2026. These placements pull from the same Merchant Center data your Shopping campaigns use. A stronger feed doesn't just improve organic AI visibility - it improves ad quality on AI surfaces simultaneously.
Performance Max is worth addressing directly. If you're running Performance Max campaigns, your product feed is the primary input. Google's AI distributes your products across Shopping, Display, YouTube, Discover, and AI Mode based on what it can read from your data. A weak feed constrains Performance Max performance across every surface it touches.
The actions that compound across all surfaces:
- Feed audit: Pull a product-level report and identify products with missing GTINs, incomplete titles, or disapprovals - these cost you visibility in Shopping, AI Overviews, and ChatGPT simultaneously
- Title rewrites: Move from generic internal naming to structured descriptive titles (Brand + Type + Differentiator + Variant) - highest-leverage feed improvement for AI surface visibility
- Disapproval fix: Each disapproval removes that product from all Google surfaces - fix within 48 hours of identification
- Review strategy: Build a consistent post-purchase review request process - reviews feed AI surface ranking signals across multiple platforms
- Loyalty activation: If you offer loyalty pricing or exclusive delivery, activate Merchant Center loyalty annotations
- GTIN completion: Audit for products missing barcodes and complete them - Google uses GTINs to match listings to third-party review data
The advertisers who will capture the most value from Shopping AI aren't necessarily those with the biggest budgets. They're the ones whose product data is accurate, complete, and structured well enough for AI systems to confidently recommend them. For more on how feed quality affects Shopping performance, see product feed quality and Shopping ROAS.
Where This Is Heading
Agentic commerce is the end state. Not a vision for 2030 - an active development happening right now.
Google's Universal Commerce Protocol isn't a roadmap item. It's a live open standard, already integrated with Shopify, Etsy, Wayfair, Target, and Walmart. The design intent is clear: AI agents execute purchases on behalf of shoppers, not just recommend them. A shopper tells Gemini "find me the best trail running shoes in size 11 under $180 and order them" - and Gemini completes the transaction. Target and Walmart already have this working.
What this means for your business: the "product page visit" may become optional for a growing share of transactions. An AI agent will read your product data, check pricing and availability, process the order - without a human seeing your homepage. That's not a negative. It's a higher-quality transaction from a better-qualified buyer. But it means the quality of your data is doing more of the work your marketing used to do.
The take: Shopping AI doesn't kill paid advertising. It raises the floor on what a "qualified visitor" means. When an AI agent sends someone to your site, or completes a purchase on their behalf, they've been pre-qualified through a conversational filter more rigorous than a keyword match. The advertisers who lose are those who've been relying on volume over data quality.
The asymmetry right now: most advertisers are still optimising for 2023. Feed quality, GTIN completion, structured data, and AI surface readiness are all under-invested. The window to build that advantage before the rest of the field catches up is open - but it won't stay open indefinitely.
Shopping AI doesn't change what a good product is. It changes how easily a machine can identify one. The businesses with clean, trusted, machine-readable data today will be visible in more places, to better-qualified buyers, at higher conversion rates than those who treat their feed as a back-office chore.
That's the shift. The practical starting point is the same place it's always been - your Merchant Center feed.
- ChatGPT Shopping is here and it's changing eCommerce SEO rules - Search Engine Land
- Google AI Overviews Now Appear on 14% of Shopping Queries - ALM Corp
- New tech and tools for retailers in an agentic shopping era - Google Blog
- Google outlines AI-powered, agent-driven future for shopping and ads in 2026 - Search Engine Land
- Product Feed Spec for Agentic Commerce - OpenAI
- Google AI Shopping Features: How to Maximise Your Visibility - Shopify