Skip to main content

Why a prompt returned no results or no citations

Four reasons a check comes back empty, how to tell which one you have, and what to do about each.

Symptom: a prompt was checked, but the answer contains no brands, no citations, or neither.

Four causes, in rough order of likelihood. Each looks slightly different, so you can usually identify yours by reading the recorded answer.

1. The assistant answered from memory instead of searching

Most common by far.

AI assistants decide for themselves whether to search the web for a given question. When they answer from what the model already knows, there's nothing to cite — so you get a fluent, source-free paragraph.

How to spot it: the answer reads as a competent general explanation with no links, and often no specific product names. The Searches page reports this directly, as the share of answers answered without searching — so you can see how much of a prompt's history it affects rather than judging from one run. See What the AIs searched for.

What to do: rephrase toward a recommendation. "What is a carbon-plated running shoe?" gets answered from memory. "Which carbon-plated running shoes are worth buying?" pushes the assistant to look things up, because current products and prices aren't reliably in its memory. What makes a good prompt to track covers the phrasings that work.

2. The question doesn't invite brand answers

Some questions have no brands in the honest answer. "How long should a marathon training plan be?" is a real question a runner asks, and the correct answer mentions weeks, not shoes.

How to spot it: citations may be present, but no brand — yours or anyone's — appears. If competitors aren't being named either, the question is the issue, not your visibility.

What to do: keep it if you want to know who owns the informational territory; replace it if you're measuring competitive position. This is the distinction between a prompt you're losing and a prompt nobody wins.

3. The prompt is beyond your capacity

When you're at your limit, the newest prompts pause while the oldest keep running. A paused prompt produces no data at all.

How to spot it: there's no recorded answer whatsoever, rather than an answer that lacked brands. Check your usage against your allowance on the prompts page.

What to do: archive prompts you no longer need, or add capacity. See How limits work.

4. The platform isn't on your plan

If a prompt is set up for a platform your plan doesn't include, that particular check won't run — while the same prompt on your included platforms runs normally.

How to spot it: data exists for some platforms on the prompt and not others. The platform picker shows which are available to you.

Quick diagnosis

Read the recorded answer for the run:

  • No answer recorded at all → cause 3 or 4

  • An answer with no links → cause 1, rephrase it

  • An answer with links but no brands anywhere → cause 2, decide if the prompt is worth keeping

  • An answer with links and competitors but not you → none of these. The tracking worked and you're genuinely losing that question, which is the signal MentionBird exists to give you. Ranking advice is the next step.

Did this answer your question?