brand_insights: what buyers ask AI about your brand
brand_insights is the read-only tool of the Data AI Search MCP server. It returns a brand's Insights: the real questions its buyers ask AI, whether the AI answers name the brand, which competitors and sites they name instead, plus keywords and customer segments.
It's where to start. Called without a brand, it lists the workspace's brands, and the other two tools take a brand name from that list.
What it can read and do
Every call names a brand and either a view (what to read) or, on the two audit tools, an action (what to change). With neither, the tool returns its default view.
brands- The workspace's brands, each with its website, category, market, language, the status of its Insights and whether it has an AI Visibility audit and a Website Audit. The default when no brand is given.
summaryDefault- The brand, what it has set up, and its Insights headline: mention rate, top competitors and sources. The default when a brand is given.
overview- The full Insights overview: metrics, changes since the last refresh and highlights.
questions- The questions the brand's customers really ask: discovery questions, which never name a brand, and consideration or compare questions, with stage and trend.
answers- How AI answers those questions: whether the brand is named, its position, and the competitors and sites cited.
keywords- The keywords that run through the brand's questions.
segments- The customer segments the questions come from.
search- The brand's questions closest in meaning to text, in any language, optionally only those seen in the last days.
What a call costs
- Every call costs 1 credit from the workspace the API key belongs to.
- brand_insights never starts a run, so a call never costs more than that 1 credit.
- Reading Insights doesn't refresh them. They refresh on their own, weekly on paid plans.
Input parameters
Every parameter the tool's input schema accepts. Your assistant fills these in from what you ask; you never type them yourself.
| Parameter | What it does |
|---|---|
brand | The brand's name, case-insensitive. Leave it out to list the brands. |
view | What to read: brands, summary, overview, questions, answers, keywords, segments or search. |
text | view search: what to find similar questions for. Any language works. |
days | view search: only questions seen in the last N days. |
limit | view search: how many results to return. |
Questions that use this tool
Ask in plain language and the assistant picks the view or action. These make good first questions.
- “Which brands can you see in Data AI Search?”
- “What are [brand]'s customers asking AI about [category] this month?”
- “Which competitors does AI name instead of [brand]?”
- “Find [brand]'s questions about price or delivery, in any language.”
- “Which customer segments ask the most about [brand]'s category?”
Questions about this tool
Can brand_insights change anything?
No. It only reads. Changes go through ai_visibility_audit and website_audit.
Why does it need a brand name?
Every view except brands is about one brand. If the name doesn't match, the tool replies with the names the workspace has, so the assistant can pick the right one.
Does search work across languages?
Yes. It matches by meaning, so an English query also finds Indonesian, Malay or Vietnamese questions on the same topic.
Is it safe to auto-approve?
It's the only tool that never changes anything, so allowing it without a prompt is reasonable. Each call still costs 1 credit.