[Sightlabs]
Docs / Alerts

Alerts

Alerts notify you when something worth acting on happens in your scan results. This page explains what triggers an alert, how to receive them, and how to tune the thresholds to reduce noise.

PR risk alerts

A PR risk alert fires when the AI pipeline detects a pattern that suggests your reputation may be under pressure. There are three trigger types:

Negative spike

The share of negative mentions jumps by 5 or more percentage points compared to the previous scan run. This is the most common trigger. It catches complaints that multiply quickly rather than individual one-off posts.

Viral thread velocity

A single negative mention gathers a large number of replies or engagements within a short window. The exact threshold depends on baseline engagement for your query. Sightlabs flags threads that are growing faster than your historical average.

Sentiment inversion

Your positive and negative lines cross on the trend chart. Positive sentiment drops below negative for the first time in the analysis window. This is a lagging signal but indicates a sustained shift rather than a one-scan anomaly.

Note
Alert classification is explained in detail in the Intelligence docs. PR risk is one of the action item types the AI pipeline produces on every scan.

Notification channels

Currently, alerts are delivered by email to the address on your account. The email includes the alert type, the scan that triggered it, and a link directly to the result.

Tip
Slack and webhook delivery are on the roadmap. Team plan subscribers will get custom alert webhooks when that feature ships.

Alert rules and thresholds

Each tracked entity has its own alert configuration. You can adjust thresholds per entity from the entity settings panel in the dashboard.

Negative spike thresholddefault: 5 percentage points

Raise this if your brand regularly sees high negative volume and you only want to hear about significant changes. Lower it if you want early warnings on smaller shifts.

Minimum mention volumedefault: 10 mentions

Alerts will not fire if the triggering scan returned fewer than this number of total mentions. Prevents false positives from low-data scans.

Alert frequency

Sightlabs will not send more than one alert per entity per scan run. If multiple trigger conditions fire on the same run, you receive a single email that lists all of them.

For scheduled scans (Pro and Team plans), the alert cadence matches your scan cadence. Daily scheduled scans can produce at most one alert per day per entity. Hourly scans can produce at most one per hour.

Muting alerts

If you know a negative spike is expected (a planned product change, a controversy you are already handling, a price increase announcement), you can mute alerts for a specific entity for a set period. Muting stops email delivery but does not affect scan results or the analytics trend.

To mute, open the entity in your dashboard and set a mute window of 24 hours, 72 hours, or 7 days. The mute expires automatically. You can unmute early by returning to the entity settings.

See also

Intelligence · how PR risk is classifiedAnalytics · sentiment trend charts