Does the type of question asked change how often companies get named

A reasonable hypothesis going into this study was that engines would name companies more readily for some kinds of prompts than others. A commercial-investigation prompt, such as asking which vendors to compare before a purchase, seems like it should invite more company names than a purely informational prompt asking how something works. This finding tests that hypothesis against the data.

The study classified prompts into four intent types: commercial-investigation, informational, navigational, and transactional, and measured the appearance rate, the share of answers in which at least one tracked company was named, separately for each. Transactional prompts were by far the largest category, with 124,338 answers and 47,628 appearances, an appearance rate of 38.3%. Commercial-investigation prompts produced 21,885 answers with 8,457 appearances, a rate of 38.6%. Informational prompts produced 11,541 answers with 4,469 appearances, a rate of 38.7%. Navigational prompts produced 12,693 answers with 4,687 appearances, a rate of 36.9%.

The four rates cluster tightly, from 36.9% at the low end to 38.7% at the high end. That is a narrow spread given how different these prompt types are in intent and phrasing. The initial hypothesis, that commercial or transactional intent would clearly outperform informational or navigational intent in prompting engines to name companies, is not supported by this data. If anything, informational prompts had a very slightly higher rate than commercial-investigation prompts, which runs opposite to the intuitive expectation.

Why might prompt intent matter so little to the appearance rate? One explanation is that the engines' decision to name a company is driven mainly by whether the prompt's topic maps onto a category where the engine has confident knowledge of specific players, regardless of whether the question is framed as 'what is X' or 'which company should I use for X.' If an engine has strong signal about companies in, say, commercial cleaning services, it may name them whether the prompt is informational or transactional in framing, because the underlying topic pulls in the same candidate companies either way.

A second explanation is that the four-way classification of intent may not cleanly separate prompts by how much they actually invite company names in practice, even if the categories are meaningfully different from a marketing or SEO perspective. A prompt labeled informational, if it happens to ask about a specific service category, may still naturally invite a company name as an example, narrowing the gap with more explicitly commercial prompts.

What this means for a company deciding where to focus content or outreach efforts: this finding does not support prioritizing commercial or transactional content over informational content on the theory that only commercial framing gets companies named by AI engines. The appearance-rate data suggest the topic and category matter more than the intent framing of the question. A company should focus on being a strong, specific answer within its category rather than assuming that only bottom-of-funnel, comparison-style prompts are worth optimizing for.