What is a chat feature?

A chat feature is a structured component inside an AI answer, as opposed to plain prose. Product cards, comparison tables, inline citations, image grids, maps, follow-up prompt chips, and sponsored slots are all chat features. They are the AI answer equivalent of SERP features on a Google results page.

How is this different from brand mention tracking?

Mention tracking answers whether you appear in an answer. Chat Features answers where and in what form. A brand named once in a paragraph of prose and a brand occupying the first product card are both mentions, and they perform nothing alike. Chat Features records which component you occupy and where that component sits.

Which AI platforms are covered?

ChatGPT, Gemini, Perplexity, Google AI Mode, Google AI Overviews, Copilot, Grok, DeepSeek, and Meta AI. Coverage per feature type varies because platforms render different components, and the tracker records which platform supported which feature on every capture.

Introducing Chat Features

Q: Is Chat Features available now?

Yes. It is part of DemandMetrics for Gen AI, available in the application and through the API, and in the Search Intelligence BigQuery warehouse for customers with that add-on.

You’ve heard of SERP Features, but have you heard of Chat Features?

Ask an AI platform a commercial question today and you often will get more than a simple text response with some links. You’ll get rich results like a product carousel, a comparison table, a set of cited sources, and a row of follow-up prompts. Every one of those is a component the platform chose to render, in a slot it chose to render it in.

Sounds familiar, right?

Today we are shipping Chat Features, which tracks those components the way SERP feature tracking has covered Google results pages for years.

It is the first feature-level analytics for AI answers available anywhere.

Mentions and citations were a good first start, but they aren’t enough anymore

For two years the industry has measured whether a brand appears in AI answers and citation links.

Search teams solved this problem on Google a long time ago. Knowing you rank third means little without knowing whether an AI Overview, four ads, and a People Also Ask block sit above you. That is why Visual Rank exists. AI answers now need the same treatment.

What we track

Chat Features records, for every captured answer:

Which components rendered, across 24 feature types (and growing) including maps, shopping features, product cards, images, video, and more
What slot each occupied, and in what order down the answer
How much height each took up, so displacement is measured rather than guessed
Which brands sat inside each component, including your competitors
What changed since yesterday, with alerting on the features you care about

Why are we launching this now?

Chat Features are taking over generative AI responses the same way that SERP Features took over the SERPs. Simply tracking mentions and citations without this context is quickly becoming meaningless.

Feature support also varies more than people expect. The same question produces a shopping carousel on one platform, a citation list on another, and plain prose on a third, and that changes without notice. Measuring it per platform is the only way to know which surfaces are worth optimizing for.

Available today

Chat Features is part of DemandMetrics for Gen AI, in the application and through the API, and in the Search Intelligence BigQuery warehouse for customers on that add-on. It joins the same dataset as SERP feature tracking, so a single view can compare how a topic behaves on Google against how it behaves in AI answers.

See the full breakdown, or talk to your Account Manager about turning it on.