Answers

Why did my AI visibility drop?

A falling score can be alarming, but it's rarely random — the cause is almost always one of a handful of specific, checkable things.

A drop in AI visibility usually comes from one of four things: a competitor published new content an AI engine now prefers citing, the underlying model was updated and changed which sources it favors, your own content or entity data went stale or inconsistent across the web, or the query set you're tracking simply shifted. Checking which one applies starts with looking at what changed for your competitors, not just your own numbers.

  • A visibility drop is rarely something you actively broke — it's usually something that changed elsewhere.
  • Competitor activity is the most common cause: check what they published or got cited for right before your drop.
  • AI engines update periodically, which can shift which sources they prefer without any change on your end.
  • Inconsistent entity data — your name, description, and details differing across the sites AI engines draw from — is a frequently overlooked cause.

The four most common causes, in order of likelihood

  • A competitor published something new

    Check whether a competitor recently earned new coverage, a comparison mention, or a review that an AI engine may now be citing instead of, or alongside, you.

  • The AI engine changed how it sources answers

    Model and retrieval updates happen periodically and can shift which sources get preferred — not something you can fix directly, but worth ruling out before assuming the problem is your content.

  • Your entity data went inconsistent or stale

    If your name, description, or details differ across the third-party sites an engine draws from, that inconsistency can quietly erode how confidently it cites you.

  • Your tracked query set shifted

    If you changed your prompt list, competitor list, or date range between checks, part of an apparent drop can just be a different measurement, not a real change.

How to figure out which one is yours

  1. 1

    Compare your drop against competitor movement, not just your own trend

    If a competitor's mentions rose right as yours fell, that's a strong signal of what actually happened.

  2. 2

    Re-run the exact same prompt set and date range you used before

    A different prompt list or competitor set between checks can look like a drop when it's really just a different measurement.

  3. 3

    Check whether your entity details are consistent across the web

    Look at how your brand is named and described on the third-party sites AI engines already cite for your category.

  4. 4

    Look for anything that changed in your own content around the same window

    A page taken down, rewritten, or gone stale can explain a drop just as easily as anything external.

Why one drop isn't a verdict

AI-generated answers carry some inherent volatility — the same prompt run twice in the same week can return a slightly different mix of brands, even with nothing having changed. A single lower reading is a data point, not a trend.

The more reliable signal is a sustained move over several weeks of consistent tracking, not a single check that happened to catch a low point. Treat one drop as a prompt to look closer, not as confirmation something is broken.

Frequently asked questions

Does a drop in AI visibility mean I did something wrong?
Not necessarily — the most common causes are external: a competitor gaining new coverage, or an AI engine updating how it sources answers. Start by ruling those out before assuming the problem is on your side.
How often do AI engines change how they source answers?
There's no fixed schedule, and providers don't always announce these changes — which is why comparing your trend against competitor movement is more useful than trying to track model updates directly.
What is "entity data" and why does it affect this?
It's how your brand's name, description, and key details are represented across the third-party sites an AI engine draws from. If those details are inconsistent or outdated, the model has a harder time citing you confidently.
How long should I wait before treating a drop as a real trend?
Several weeks of consistent tracking on the same prompt set, rather than a single re-check — a one-off dip is common and often self-corrects.

See exactly what changed before your next visibility drop

Track competitor movement, sentiment, and citation sources alongside your own score — so a drop comes with a reason, not just a number.

Related answers