AI Visibility Trends & Alerts

Reading the trend chart

The chart plots the percentage of runs in which you were named or cited, one line per engine. Look for direction rather than day-to-day wobble: a single low day can be nothing more than natural variation between answers, while a steady slide over two weeks deserves attention.

Because each point is built from runs in the recent window, more frequent checks and more runs per prompt make the line smoother. If your chart looks jumpy, that is a hint to run each prompt a few more times, as described in AI prompt tracking.

Drop alerts

You do not need to watch the chart to notice a problem. When your visibility rate falls by a significant number of points compared with about a week earlier, an alert is raised for that website and engine. It appears alongside your other alerts in the monitoring and alerts inbox, and can be sent out through webhooks like any other alert.

• Enough data first: an alert only fires when there are enough successful runs to trust the comparison.
• Once a week at most: repeated alerts for the same website and engine are combined, so you are not buried in notifications.
• Failures are ignored: runs that failed are left out entirely, so a provider hiccup can never look like a collapse in visibility.

Working out what changed

When an alert arrives, start with the answer history. Read a few recent answers for your key prompts and compare them with older ones: has a competitor moved up, has a new source appeared, or has the wording about you changed? The share of voice and competitor view will show whether a rival is gaining while you stand still.

Then check your own website. A recent redesign, a removed page or a change in your technical SEO can all affect whether engines find and trust your content. Because everything sits in one account, you can move from the alert to the crawl results in a couple of clicks.

Keeping a record

Snapshots are kept long after the individual answers have aged out, so you can show a client the story of several months rather than a single moment. Bring that history into your regular client reporting and the trend becomes part of the monthly conversation instead of a surprise.

Setting realistic expectations

AI visibility is a slower-moving measure than most people expect. Improvements you make to your content or reputation usually show up over weeks, not days, and the numbers will always carry some noise. Judge progress over a month or more, compare like with like across the same prompts and engines, and treat an alert as a prompt to investigate rather than proof that something is broken. That habit keeps you calm when a figure wobbles and quick to act when it genuinely slides. Note down the date of any big content release or site change so you can line it up with the chart later and learn which of your efforts actually move the needle.

Frequently asked questions