Does the kind of question a buyer asks change how often a company shows up?

AI assistants get asked business questions in different registers. Sometimes a user wants to compare options directly ("best providers for X"), sometimes they are researching a topic before they have decided to buy anything, sometimes they are looking for a specific company by name, and sometimes they are ready to transact. This study tagged its 14,951 prompts into four categories, commercial-investigation, informational, navigational, and transactional, and measured how often companies appeared in answers to each.

The appearance rates turn out to be close to each other. Informational prompts produced the highest appearance rate, 38.8% of 89,931 answers, with 34,886 total appearances. Commercial-investigation prompts, the ones that read most like direct buyer comparisons, came in close behind at 38.6% across 122,949 answers and 47,496 appearances. Transactional prompts sat at 37.9%, the largest single bucket by volume with 700,575 answers and 265,811 appearances. Navigational prompts, where a user is often looking for something specific, had the lowest rate at 37.4% across 100,170 answers and 37,471 appearances.

The spread across all four categories is narrow, roughly a point and a half between the highest and lowest. That is a notable finding in its own right, because it means the type of question a buyer asks does not strongly change the odds that a company gets mentioned at all. A company is not meaningfully safer or more exposed depending on whether buyers are asking broad research questions or narrow, ready-to-buy ones. Whatever is driving which companies get named, it appears to operate similarly regardless of where in the buying journey the prompt sits.

This is somewhat counterintuitive. One might expect commercial-investigation prompts, phrased explicitly as comparisons, to produce longer and more inclusive lists, and navigational prompts, which often name or imply a specific company already, to produce the narrowest and most selective lists. The data shows navigational prompts with the lowest appearance rate, consistent with that logic, but the gap to the highest category, informational, is small, about a point and a half, not the large gap that logic might predict if navigational intent sharply filtered which companies get surfaced.

A plausible explanation is that the engines are not adjusting their underlying candidate pool much based on prompt phrasing. The study's own category definitions distinguish intent, but the mechanism each engine uses to decide which companies to surface may be more sensitive to the category being asked about, such as mobile app development versus restaurant software, than to the register of the question within that category. That would be consistent with the much larger spread seen across engines in other findings in this study, such as the gap between ChatGPT's and Gemini's overall visibility rates, compared to the narrow spread seen here across prompt types.

What this means for a company trying to improve its AI visibility is that there is probably little to gain from writing content narrowly targeted at one stage of the buyer journey in the hope of catching a more favorable prompt type. A company invisible to informational prompts is, on this evidence, about as likely to be invisible to transactional ones, give or take a small margin. The more productive lever, based on the other findings in this study, is likely to be engine coverage and category fit, not prompt-stage targeting.