Get ready for product discovery in the age of AI.
Over the past six months, the infrastructure for AI-powered commerce (or “agentic commerce”) has been built out at remarkable speed:
- Google launched Shopping in AI Mode and announced the Universal Commerce Protocol in January 2026, co-developed with Shopify, Walmart, Target, and Etsy.
- OpenAI launched the Agentic Commerce Protocol with Stripe in September 2025, enabling purchases directly within ChatGPT.
- Microsoft followed with Copilot Checkout in January 2026.
- Shopify is positioning itself as the connective layer across all platforms.
As we see agentic commerce development picking up steam, advertisers want to know: what does this mean for how my products get discovered?
At Shoparize, we believe this is another step in the direction we’ve already seen over the last few years: structured data quality is becoming the cornerstone for product discovery in e-commerce. It might just be your competitive edge in the age of agentic commerce.
In this article, we’ll explain exactly how you can get ahead as a merchant.
Expanding product attributes to be AI-ready
Traditional Google Shopping matches search terms to products. A search for “blue running shoes” triggers ads bidding on that term. The merchant with the highest bid and best Quality Score wins the auction.
AI Mode works differently. Shoppers using AI Mode are asking questions that are 2 to 3 times longer than traditional search queries. Instead of “running shoes,” they’re typing: “I need wireless headphones under €200 with active noise cancellation for my daily train commute, comfortable enough to wear for 2 hours, and compatible with both my iPhone and laptop.”
Behind the scenes, AI Mode uses semantic modelling to match the purpose of the query with product attributes in the Shopping Graph. It interprets intent and matches against structured product data. A query like “lightweight waterproof jacket for cycling in the rain under €150” requires the AI to find those specific attributes in your feed.
No attribute, no match. No match, no visibility.
One of the biggest upgrades is query fan-out, where Gemini automatically breaks a single query into micro-questions to broaden its understanding. A search for “best waterproof travel bags for weekend trips” might generate dozens of parallel queries: waterproof materials, bag dimensions for carry-on, durability ratings, zipper quality, and organisational pockets. Each micro-query hits different segments of the Shopping Graph, pulling in products that match multiple overlapping signals.
This is why the data attributes provided to optimise for AI discovery are in need for expanding.
Improving “feed quality” for AI readability
Google is rolling out dozens of new data attributes in Merchant Center designed for easy discovery in the conversational commerce era. These new attributes go beyond traditional keywords to include things like answers to common product questions, compatible accessories or substitutes.
In AI shopping, there’s no second page of results. As a result, it behaves quite binary: you’re either matched to the intent, or you’re invisible. Missing or inconsistent data can quietly disqualify you from entire categories of queries. The retailers who win visibility in AI Mode will be the ones with:
1. Attribute depth beyond basics: Colour and size aren’t enough. Material composition, use case, compatibility, and care instructions. Details a knowledgeable shop assistant would know. Product description has gained immense value. While titles are concise, the description allows for up to 5,000 characters of rich, keyword-heavy text. Google’s AI will parse this to understand nuanced features, use cases, and technical specifications that answer complex shopper questions.
2. Titles and descriptions written for understanding: A title like “Men’s Waterproof Gore-Tex Hiking Jacket” is more relevant than just “Men’s Jacket.” The old SEO playbook, stuffing titles with every possible keyword variant, works against you here. AI needs to understand what your product actually is and does. Write titles that describe, not titles that game algorithms.
3. Consistent data across surfaces: Structured markup should be consistent with the feed. Merchant Center, website content, and structured data need to align. AI cross-references sources. Inconsistencies confuse matching. The Shopping Graph refreshes over 2 billion listings per hour. If your feed lags behind your actual inventory, you’re creating friction.
4. High-quality images: Virtual try-on (live in the UK) and visual search rely on image quality. Google Lens processes billions of visual queries monthly. AI increasingly “reads” images using vision models that analyse colours, textures, shapes, and packaging.
What can merchants do right now?
There’s no separate integration required for AI Mode. Your current Merchant Center feed is what the AI uses. If and how AI uses your feeds is up to the actions you take today:
1. Audit attribute completeness: Pull your feed. Check coverage on material, use case, compatibility, care instructions, and technical specifications. Flag gaps. Prioritise high-margin and high-volume SKUs first. Think about the questions a shop assistant would answer: “Is this compatible with X?” “What’s the difference between this and that?” “Will this work for my situation?” Your feed should contain those answers.
2. Rewrite titles for meaning: Pick your top 100 SKUs. Rewrite titles that read like keyword strings into titles that read like product descriptions. Before: “Men’s Jacket Waterproof Jacket Rain Jacket Cycling” After: “Men’s Waterproof Cycling Jacket. Lightweight, Breathable, Reflective Details”.
3. Expand descriptions to 5,000 characters: Front-load the most important information and keywords within the first 150-180 characters. Then use the remaining space to answer common questions, explain use cases, and provide technical detail. This is where AI finds the information it needs to match your product to complex queries.
4. Align your sources: Spot-check that Merchant Center data matches your website content and structured data. Schema markup on your site should reflect what’s in your feed. Inconsistencies create uncertainty for AI systems.
5. Set up proper refresh cycles: Real-time is ideal. If not possible, prioritise accuracy in pricing and stock levels. Nothing kills AI visibility faster than recommending an out-of-stock product or showing the wrong price.
The bottom line
Brands that treat product data as a strategic asset will win the AI shelf space. Your feed is no longer just a data file you send to Google. It has become your storefront in an AI-first discovery layer that’s already serving 75 million users daily.
Want to speak to our experts on Google Shopping Ads, Merchant Center and data feeds? Get in touch today to find out how Shoparize can help you win product discovery: https://partner.shoparize.com/en/contact
