Automated insights are AI-discovered patterns and anomalies in your data. Learn how to read and act on them.
What Are Automated Insights?
Insights are AI observations about your data. Each one falls into one of four types:
| Type | Example |
|---|---|
| Anomaly | "Traffic dropped 35% compared to last week" |
| Trend | "Organic traffic up 25% over 4 weeks" |
| Recommendation | "Mobile converts 40% less than desktop" |
| Prediction | "Visitors are on track to exceed last month" |
Where to Find Insights
Insights live in two places:
- Per website: open your website's dashboard and select the Insights tab (under Behavior in the sidebar).
- Across your team: the Insights page in the main sidebar aggregates the most important insights across every website in your workspace.
Insight Anatomy
Insight Card Structure
Each insight card shows:
| Element | Purpose |
|---|---|
| Type | Anomaly, trend, recommendation, or prediction |
| Title | Brief description |
| Severity | Importance level |
| Description | What changed and why it matters |
| Detected | When the insight was generated |
Open a card to see the full detail view, which adds the supporting chart, the current value, the expected value, and the deviation.
Example Insight
ANOMALY: Traffic Drop Detected
Your website traffic dropped 35% compared to
the same day last week.
Current: 780 visitors
Expected: 1,200 visitors
Deviation: -35%
Detected: Today, 2:00 PM
Insight Types
Anomalies
Unusual deviations from the expected range:
| Anomaly | Meaning |
|---|---|
| Traffic spike | Sudden increase |
| Traffic drop | Sudden decrease |
| Conversion change | Rate significantly different |
| Error surge | More errors than usual |
Trends
Patterns developing over time:
| Trend | Example |
|---|---|
| Growing | "Organic traffic up 5% weekly" |
| Declining | "Bounce rate increasing steadily" |
| Seasonal | "Traffic peaks on weekends" |
Recommendations
Potential improvements found in your data:
| Recommendation | Example |
|---|---|
| Conversion gap | "Mobile converts 40% less than desktop" |
| Traffic source | "Social visitors are more engaged" |
| Content | "Blog drives 3x more signups" |
Predictions
Where a metric is likely heading based on recent activity, for example "Visitors are on track to exceed last month's total."
Insight Severity
Severity Levels
| Level | Color | Meaning |
|---|---|---|
| Critical | Red | Immediate attention needed |
| Warning | Orange | Should investigate soon |
| Info | Blue | Good to know |
Critical Insights
Require immediate action:
- Major traffic drops
- Conversion rate crashes
- Error surges
Warning Insights
Investigate within a day:
- Moderate changes
- Developing trends
- Potential issues
Info Insights
Review when convenient:
- Minor changes
- Context updates
- Background patterns
Reading Insight Data
When you open an insight, the detail view shows the numbers behind it:
Conversion Rate Change
Current: 2.1%
Expected: 3.4%
Deviation: -38%
Current vs. Expected
The AI compares the current value against an expected value derived from your recent history. The deviation is how far the current value sits from that expected baseline. A historical chart plots the metric over time so you can see the change in context.
Acting on Insights
From an insight card or its detail view you can:
| Action | What it does |
|---|---|
| Open | View the full detail, chart, and supporting metrics |
| Acknowledge | Mark the insight as reviewed and clear it from the active list |
| Dismiss | Remove an insight you don't find useful |
| Copy | Copy the insight summary to your clipboard to share |
Acknowledged insights are hidden from the active list by default. Use the Show acknowledged toggle to bring them back into view.
Managing Insights
Insight Feed
To work through your insights:
- Open the Insights tab on a website's dashboard, or the team-wide Insights page.
- Browse the list, newest first.
- Filter or search to narrow it down.
Filtering Options
The toolbar above the list lets you filter and search:
| Filter | Options |
|---|---|
| Type | Anomaly, Trend, Recommendation, Prediction |
| Severity | Critical, Warning, Info |
| Acknowledged | Show or hide acknowledged insights |
| Search | Match against the title, description, or metric |
Acknowledging Insights
Why Acknowledge?
- Marks the insight as reviewed
- Clears it from the active list
- Keeps your feed focused on what still needs attention
How to Acknowledge
- Read the insight
- Take any needed action
- Click Acknowledge
Best Practices
Daily Review
- Check critical insights
- Review new anomalies
- Acknowledge actioned items
Weekly Review
- Look at trends
- Review recommendations
- Check developing patterns
Acting on Insights
- Read - Understand the insight
- Verify - Confirm in your reports
- Investigate - Find the root cause
- Act - Take appropriate action
- Acknowledge - Close the insight
Building Habits
- Include the Insights tab in your daily routine
- Share notable findings with your team
Troubleshooting
Too Many Insights
If overwhelmed:
- Filter by severity
- Acknowledge older items to clear the list
- Focus on critical only
Missing Insights
If you're not seeing expected insights:
- Check the type and severity filters
- Make sure acknowledged insights aren't hidden when you expect to see them
- Verify your plan includes AI Insights (Pro and above)
- Contact support
Irrelevant Insights
If an insight isn't useful, dismiss it to remove it from your feed.