Quattr

Quattr MCP gives an AI assistant read-only access to one company's Quattr data: how AI answer engines cite the brand, how its pages perform in Google search, and how it compares with tracked competitors. Use it to investigate changes, find competitive gaps, prioritize keywords and pages, and pull numbers for reports. Answer from this data, not from general knowledge.

What it covers

AI answer visibility: citations and share of voice in AI answer engines such as ChatGPT, Perplexity and Gemini, the pages that get cited, and the prompts where competitors appear instead. Each tracker measures its own set of tracked prompts.

Google AI Overviews: how often the site appears in the AI Overview on Google's results page. This is a separate measure from AI answer citations, so always say which one a number refers to.

Organic search: Search Console clicks, impressions, click-through rate and average position, by page, query, country, device, intent or content category.

Competitors: search market share and rankings against tracked competitors, and the keywords where they outrank the company.

Technical health: AI crawler activity from server logs, indexability, crawl efficiency and Core Web Vitals.

Search to revenue: web analytics from GA4 or Adobe Analytics, conversions under the company's own goal names, and Google Ads performance, where that data is connected.

Common workflows

"Who gets cited instead of us in AI answers?" Confirm the tracker was collecting for the period. Compare share of voice with competitors, list the prompts where the company is missing, then show which competitor pages win them.

"Which keywords are close to page one that we should push?" Take non-branded keywords ranking just off page one and estimate the extra clicks from the site's own click-through rates by position. Present that as an estimate, not a forecast.

"Why did our organic clicks change last week?" Compare the last complete week with the one before. Check data freshness, find the pages and queries that moved most, and test whether the change is real before explaining it.

"Create our weekly search report." Cover the last complete week against the week before, label the dates, and flag any data source that is running late.

How to read the numbers

Look up any segment, AI tracker, goal or competitor set the user names before filtering by it, and use the exact match. A single competitor's domain can go straight into a market share comparison.

Date ranges include both end dates. Quattr weeks run Sunday to Saturday, and "last week" means the most recent complete one.

Market share is a modeled share of tracked search demand on a 0 to 100 scale, so 0.03 means 0.03%, not 3%. Report changes in percentage points and treat the figure as directional.

For average position, lower is better. A rising number means a falling rank.

Keep branded and non-branded results apart when the split changes the story. Default to non-branded for opportunities and say which you used.

A named segment covers only its own keywords, so its numbers don't compare with the Overall Market view.

Low AI visibility can be a tracking gap rather than a performance gap. Check the tracker's prompt coverage first. Edits to a tracker's prompts also break its trend line.

Check freshness before any claim about the last few days. Search Console runs one to two days behind, and Core Web Vitals usually refresh about every two weeks. A stalled feed is broken, not slow.

Call a change real only after a significance test, and label the result real, noise or can't tell. Correlation isn't causation, so suggest a controlled test rather than name a cause.

Limits

It reads Quattr data and never changes the website, ad campaigns or Quattr settings. It can't see a competitor's private analytics, and it doesn't promise rankings or AI citations. If the data can't answer a question, say what's missing instead of estimating. Requires a Quattr account. Request access at https://www.quattr.com/get-started.