Google Just Built Its Own AI Share-of-Voice Tool

TL;DR: Google Marketing Live 2026 shipped "AI Performance Insights" inside Merchant Center, a native dashboard that benchmarks your share of voice against competitors across AI Mode, Gemini, and AI Overviews. Google also rolled out "Conversational Attributes," a new product-data field type built for conversational queries, plus BNPL inside Google Pay for Universal Cart and UCP coverage for hotels and food delivery. The bar for measuring AI visibility on Google just moved from "specialty tool" to "free, built-in." Cross-platform measurement is what's left as the real moat.

For most of 2025 and the first half of 2026, every brand that wanted to know how it appeared on Google's AI surfaces had to buy a third-party tool, build something in-house, or hire an agency to run manual prompt audits. As of yesterday's Google Marketing Live keynote, that's no longer true. Google put the dashboard inside Merchant Center, made it free, and gave it a name: AI Performance Insights.

That single change reframes the next 90 days of brand-side AI readiness work. If you sell into the US, Canada, Australia, India, or New Zealand, you're about to have a default benchmark for Google AI visibility whether you wanted one or not. The question shifts from "are we measuring this at all" to "are we measuring it everywhere agents shop."

What Google actually shipped at GML 2026

Six product announcements landed yesterday. Three of them are the load-bearing ones for any retailer or brand thinking about AI shopping readiness.

AI Performance Insights is a new reporting view inside Merchant Center that compares your brand's AI share of voice against similar competitors across AI Mode, Gemini, and AI Overviews. It's rolling out "in the coming months" in five markets. This is Google productizing AI share-of-voice measurement natively, for free, inside the same console merchants already log into to manage feeds.

Conversational Attributes are new product-data fields in Merchant Center that brands fill in to optimize for conversational shopping queries. Rolling out globally. Where traditional product feeds were optimized for keyword match against query strings, Conversational Attributes are optimized for the fuller, intent-rich queries shoppers type into conversational commerce surfaces. Google is, in effect, declaring a new field type that the legacy feed wasn't built for.

BNPL embedded in Google Pay for Universal Cart. Affirm and Klarna are now wired natively into the Google Pay flow that powers Universal Cart checkout. BNPL used to be a separate redirect; now it's part of the agentic checkout step. With Klarna also live inside ChatGPT as of this week, BNPL is natively present in both major agentic checkout flows.

Google also expanded Universal Cart into hotels and food delivery (Booking, Expedia, and food delivery via Google Maps), launched Direct Offers with AI-generated promotion bundles, and put Gemini-powered "Ask Advisor" inside Ads, Analytics, and Merchant Center. UCP is no longer retail-only.

The headline number from GML: AI Mode now has 1 billion monthly active users, AI Overviews has 2.5 billion, and AI Mode searches run 3x longer on average than traditional Search queries. (Source: Philipp Schindler, GML 2026 keynote) Brainstorming-style queries ("ideas for X", "which X") are growing 30% faster than AI Mode searches overall.

Why "AI Performance Insights" matters more than it sounds

The product itself is a benchmark dashboard. The category implication is bigger.

Before yesterday, the question "how visible is our brand on Google AI surfaces" had a friction cost attached to it. You needed a specialty vendor, a custom internal tool, or a manual prompt audit. That friction was a moat for any vendor in the space (Paz included). Google just zeroed out that friction inside Merchant Center for five major markets. By "the coming months," every brand that ships feeds to Google will have a free default answer to "how am I doing on Google AI."

Two things happen next.

First, AI visibility moves from "specialty conversation" to default operational metric. If your team runs paid campaigns in Google Ads, they're already in Merchant Center. AI Performance Insights will be one tab over. The question shifts from "should we be measuring this" to "what's our number, and why is it lower than the competitor's." That's a moment for the AI visibility commerce conversation in the C-suite.

Second, single-platform measurement becomes free, but cross-platform measurement becomes the differentiator. Google's tool only sees Google. It doesn't tell you how you appear inside ChatGPT Shopping, Perplexity, Microsoft Copilot, or Amazon's Alexa for Shopping surface. Roughly 1 in 8 US consumers begins shopping on a GenAI chatbot now per Bain's recent benchmark; that traffic is split across platforms, not concentrated on Google. Any brand that benchmarks only on Google misses most of the picture.

Conversational Attributes is Google validating the data-readiness thesis

The most-overlooked GML announcement might be Conversational Attributes. The product itself sounds prosaic. The signal is loud.

Google is explicitly saying: the product data you wrote for keyword search isn't enough for conversational queries. There's a different field type now, and brands need to populate it.

This is the entity optimization for AI search argument made by Google itself. Brands have been hearing this from analysts and vendors for 12 months. Now it's named. The implication: every existing product feed has a new column to fill, and brands that fill it well will surface preferentially in AI Mode and Gemini conversational queries.

Our April 15 post on product data as infrastructure called the Conversational Attributes rollout when it was still a "coming" item. It's no longer coming. It's shipping globally.

What changes for retailers and brands this quarter

Three things are different on May 22 vs May 19.

First, you have a free Google-AI benchmark coming. Plan for it. AI Performance Insights will be inside Merchant Center "in the coming months" in five markets. The day it lights up for your account, somebody on your team will pull a screenshot. Decide now whether the screenshot supports a story about agentic commerce readiness or undermines it.

Second, you have a new feed field to populate. Conversational Attributes is rolling out globally. Brands that treat it as a real data project and fill it with rich, natural-language descriptions of what the product is, who it's for, and how it differs from alternatives will show up disproportionately for the 30%-faster-growing brainstorming queries. Brands that copy-paste from the existing feed will look mid.

Third, the cross-platform gap matters more, not less. If Google's giving you a free dashboard for Google, the question your CMO will ask next quarter is "what's our number on ChatGPT and Perplexity?" Without an answer, branded vs unbranded AI queries on those platforms remain a black box. That's the gap a product feed optimization program or an external AI Readiness audit has to fill.

How AI Performance Insights compares to specialty AI-visibility tools

The split is clean. Google's tool wins on default convenience and Google-surface depth. Cross-platform tools win on coverage of every other AI shopping surface that isn't Google. Most brands need both.

What to do this week

1. Map your Merchant Center accounts to the rollout markets. If you sell into the US, Canada, Australia, India, or New Zealand, you're in the first wave. Confirm your accounts are clean and your team has Merchant Center access ready to read AI Performance Insights the day it shows up.
2. Inventory your product feed against Conversational Attributes. As soon as the new fields appear in Merchant Center, you want a plan to populate them with natural-language descriptions, not a scramble. Pull your top 50 SKUs by GMV and draft conversational descriptions now.
3. Run a cross-platform baseline. Before Google's tool gives your team a Google-only number to anchor on, run your own measurement across ChatGPT, Perplexity, Gemini, and Copilot. That baseline becomes the credibility check on whatever AI Performance Insights eventually shows.
4. Update internal AI-readiness reporting templates. Anywhere your team tracks "AI visibility" as a single number, split it into "Google AI" and "everywhere else." The two metrics will diverge.
5. Brief your CMO and head of ecommerce before they see the dashboard themselves. Better to explain what Google's about to ship than to be asked to explain a low number after the fact.

FAQ

What is AI Performance Insights?
A new reporting view inside Google Merchant Center that shows your brand's share of voice across Google's AI surfaces (AI Mode, Gemini, AI Overviews) compared to similar competitors. Announced at Google Marketing Live 2026, rolling out in the US, Canada, Australia, India, and New Zealand in the coming months.

What are Conversational Attributes?
A new set of product-data fields in Merchant Center designed to optimize products for conversational shopping queries (the kind shoppers type into AI Mode and Gemini). Rolling out globally. They sit alongside traditional product attributes, not replacing them.

Does AI Performance Insights cover ChatGPT or Perplexity?
No. It covers Google's AI surfaces only. For ChatGPT, Perplexity, Microsoft Copilot, or Amazon's Alexa for Shopping, you need either a manual audit or a cross-platform AI visibility tool.

Is this the same as Google Trends or Google Search Console?
No. Trends and Search Console measure organic search behavior and traditional ranking. AI Performance Insights specifically measures share of voice inside AI-generated responses on Google AI surfaces, which behave differently from classic ranked search.

Do we need Conversational Attributes if we already have rich product descriptions on our site?
Yes. Google's AI surfaces draw primarily from Merchant Center feed data for product cards, not from your site copy. Conversational Attributes is the field type Google built to bridge that gap.

Google just made measuring your AI share of voice on Google free and default. That's a gift for brands that hadn't started measuring; it's a forcing function for everyone else. The competitive edge has moved one layer up, to whether you can see your share of voice everywhere agents shop, not just inside the Google account where the feed already lives.