One qualifier makes AI stop naming any brand you track, a third of the time
A null response is not a competitor beating you. It is an AI answer that named none of the six brands a panel tracks at all, neither the target nor any of its five named competitors. Across 6,616 real organic responses in three B2B SaaS categories, 7.4% came back null. In project management software specifically, one use-case qualifier, "for agencies," pushed that rate to 35.2%, twelve times higher than the same category's unqualified prompts.
What we counted
Every CitedWell audit panel tracks six brands per category: a target and five named competitors. A "null response" is any organic (non-branded) prompt where zero of those six names appear anywhere in the AI's answer. That is a different measurement than a mention-rate gap. A mention-rate gap means a competitor beat the target brand. A null response means the engine answered the question with a brand set outside the six the panel is watching, or did not name any specific product at all.
We re-scored every real, search-grounded response across our 270 live audit panels (project management, customer support, and HR software; ChatGPT with web search, Gemini with grounding, Perplexity Sonar, and Claude with web search) using the same matching logic our audits use to score a client, and counted how often the organic prompts in each panel came back null. That left 6,616 organic responses to measure, 488 of them null.
The rate by category
| Category | Organic responses | Null | Null rate |
|---|---|---|---|
| Project management software | 4,265 | 432 | 10.1% |
| HR software | 912 | 24 | 2.6% |
| Customer support software | 1,439 | 32 | 2.2% |
| All categories | 6,616 | 488 | 7.4% |
Project management software runs almost four times hotter than the other two categories. That gap is not spread evenly across every kind of organic prompt. It concentrates almost entirely in one prompt intent and, inside that intent, in one specific qualifier.
Best-of prompts almost never go null
| Prompt intent | Responses | Null | Null rate |
|---|---|---|---|
| Best-of ("what are the best...") | 352 | 0 | 0.0% |
| Comparison | 1,061 | 39 | 3.7% |
| Recommendation | 5,203 | 449 | 8.6% |
Across all three categories, a best-of prompt never once returned a response with zero tracked brand names, 352 responses and zero nulls. Asking an engine to list options is a strong forcing function for naming products. Comparison prompts stay fairly low too. The null problem lives almost entirely inside recommendation-intent prompts, the open "what should I use" and "can you recommend" questions, and inside that intent it is not uniform either.
The specific qualifier that breaks it
Project management panels run recommendation prompts in three forms: plain, "...for remote teams," and "...for agencies." Splitting the null rate by form shows the category-level number is really one qualifier's number.
| Prompt form (PM software) | Responses | Null | Null rate |
|---|---|---|---|
| Plain, no qualifier | 2,035 | 12 | 0.6% |
| "...for remote teams" | 1,165 | 45 | 3.9% |
| "...for agencies" | 1,065 | 375 | 35.2% |
An earlier post on this blog showed that "for agencies" and "for hiring" drove the target brand's own mention rate to exactly zero across 1,065 and 234 responses. What that post could not show is what the engine said instead. We pulled a sample of the "for agencies" null responses to find out.
Why this changes the fix
A mention-rate gap and a null response call for different responses. If a competitor is winning a prompt your brand loses, the fix is competitive: get onto the pages that competitor already owns, or build content that out-ranks it for that specific question. If a prompt goes null, that playbook does not apply, because there is no competitor to dislodge inside the tracked set. The real finding is that the buyer question itself routes to a different market: "agencies" reads to the engine as a request for agency-specific tooling, not a filtered request for the same generalist products everyone else is fighting over.
For a brand that is a generalist tool watching its "for agencies" numbers, the honest read is not "we are losing this query." It is "this query is not about us, or our tracked competitors, at all." The fix, if there is one, is deciding whether to build content that competes for the agency-specific framing on its own terms, not tuning existing pages that answer the generalist question. Treating a null response like a lost head-to-head wastes effort competing for a slice of the market the AI is not routing to your category in the first place.
See how many of your buyer-intent prompts come back null, and whether the null responses in your category are answering a market you are not in.
Get an AI Visibility Audit, $490Methodology
Data drawn from 270 live, search-grounded audit panels (project management, customer support, and HR software brands), each run across four AI engines: ChatGPT with web search, Gemini with grounding, Perplexity Sonar, and Claude with web search. We re-scored every successful (non-error) response using CitedWell's production scoring logic (engine/scorer.ts), restricted to organic (non-branded) prompts, and counted a response as null when it contained zero mentions of any of the panel's six tracked brand names, not just the target. 6,616 organic responses total, 488 null. No development-rail or fixture data is included; all responses came from live engine calls. Data collected June-July 2026.