Discoveries
Discoveries are data findings revealed by Agents when they find something noteworthy in your data. They analyze data to generate opportunities, anomalies, and recommendations that require your attention.
What are Discoveries?
Discoveries are the output of Altertable's monitoring system:
- Generated by agents: agents continuously analyze your data
- Require human review: team members approve or reject findings
- Include context: each discovery links to its source and reference
- Provide recommendations: agents explain what they found and suggest actions
Goals of Discoveries
- Detect anomalies: identify and alert on data points that deviate from historical patterns
- Reduce noise: only notify when something meaningful happens
- Enable collaboration: team members review and discuss findings
- Accelerate response: get actionable findings without manual analysis
Discovery Lifecycle
Every discovery moves through a review workflow:
| Status | Description |
|---|---|
| Needs review | Awaiting review from a team member |
| Accepted | Approved and marked as actionable |
| Rejected | Dismissed as not relevant or false positive |
Discovery Types
Discoveries can contain different types of data:
| Type | Description |
|---|---|
| Trend change | Revenue dropped 15% week-over-week |
| Anomaly | User signups spiked unexpectedly |
| Schema change | New tables detected in database |
| Segment shift | Power user cohort shrinking |
| New models | New semantic models available |
| Events detected | Web analytics events ready for analysis |
Reviewing Discoveries
When reviewing a discovery:
- Read the description and recommendation
- Check the visualization if available
- Review the source data for context
- Approve to mark as actionable or Reject to dismiss
The platform evolves with your input, automatically optimizing the discovery process based on your reviews.
Discovery Sources
Discoveries originate from various references:
- Insights: monitor individual insights for changes
- Dashboards: track entire dashboards for anomalies
- Connections: watch database connections for schema changes
- Segments: observe user segments for behavioral shifts
- Semantic models: detect new data models and tables
- Events: monitor event tracking for new activity
Learn More
- Agents: autonomous data collaborators that help you automate the work
- Dashboards: combine insights and attach agents
- Insights: charts and analyses that agents monitor