Google’s AI Search Data: A Glimpse, Not a Full Picture

Search Engine Journal reported that Google has started sharing limited data about how its AI search features handle shopping questions. The new AI Performance Insights in Merchant Center show retailers what kinds of product terms shoppers use — but not the actual queries people typed.

This is the first real look behind the curtain of Google’s AI Overviews and AI Mode. For businesses trying to understand how their products appear in these new search surfaces, it’s a step forward. But the gaps are just as important as the data: there are no click counts, no individual queries, and no way to see whether AI search actually sent a customer to your site.

Why the Missing Data Matters for Your SEO Strategy

Traditional SEO relies on exact keywords and search volumes. AI search changes the game because it answers questions in a conversation, not just a list of links. Google’s new data tells you which attributes shoppers care about — like “arch support” or “long battery life” — but not how many people asked about them.

That means you can’t optimise for a specific keyword anymore. Instead, you need to make sure your product titles, descriptions, and structured data cover the full range of features people might ask about. The data is a signal, not a map. If you treat it like a keyword list, you’ll miss the bigger picture: AI search rewards completeness, not exact matches.

For Australian businesses, this shift is especially tricky. Google’s AI models are trained on global data, so local terms and phrases might not appear in the pilot. You have to guess what your customers are actually saying, then test.

What This Means for Australian SMBs

Small and mid-sized businesses in Australia often run lean marketing teams. You don’t have time to guess which product attributes matter to AI search. The new Merchant Center pilot gives you a starting point, but only if you already have a product feed set up — and only for US-based traffic at first.

If you sell products online, this is a wake-up call. Your product data needs to be thorough. Shoppers asking AI for “waterproof hiking boots with good ankle support” won’t see your listing if your feed only says “hiking boots.” The gap between what you list and what people ask is now a direct ranking signal.

What You Can Do Now

  • Audit your Google Merchant Center product feed and fill in missing attributes like size, material, colour, and use-case features.
  • Review the AI Performance Insights tab (if available in your account) to see which product terms Google is surfacing for your category.
  • Update product titles and descriptions to include natural, conversational phrases that answer common shopping questions, not just keyword strings.
  • Monitor your share of voice in the AI report, but remember it’s calculated against competitors already in Merchant Center — not a true market share.
  • Test your own products in Google’s AI Overviews and AI Mode to see how they appear, and note any gaps in your content.

Understanding AI search data is still new territory for most businesses. MS&VG’s AI-powered SEO service helps Australian SMBs interpret these signals and adapt their product feeds and content so they show up in the answers Google’s AI is giving.