Why marketers should track branded and unbranded AI queries

Marketers assessing their presence in AI search should distinguish between prompts that name a brand and those that do not, according to HubSpot Marketing. The distinction can help teams understand whether AI answers are influencing people who already know a company or introducing it to potential customers who are still exploring their options.

A branded prompt includes a company name, a variation or common misspelling, or the name of one of its products. Someone asking an AI assistant whether Nike shoes are suitable for marathon runners, for example, is requesting information about a specific brand. The answer may affect how that person views the company while comparing options or confirming a decision.

An unbranded prompt describes a need without identifying a company. A shopper asking for the best running shoe is looking for a recommendation rather than asking about a particular manufacturer. If an AI response mentions a brand, that company enters the shopper’s consideration; if it does not, the shopper may not encounter it through that answer.

These prompt types therefore offer different perspectives on performance. Branded queries can show how a company is presented when people ask about it directly. Unbranded queries can indicate whether the company appears when people seek a product or solution without naming a provider. Focusing on only one leaves out part of the picture.

HubSpot recommends a regular tracking routine that covers both categories and cautions marketers to interpret results carefully. The distinction also has a parallel in conventional search reporting: in November 2025, Google added a Search Console filter that separates branded and unbranded queries. That filter concerns search queries and keywords; tracking prompts in AI search requires its own approach.

What this means for marketers

Build a set of prompts that includes both direct questions about your brand and broader questions about the needs it addresses. Review the answers each week, noting whether the brand appears and how it is described. Keep the two groups separate when reporting results so that strong performance among people who already know the brand does not obscure limited visibility among those still exploring their choices.