Anomaly Detection Skill

This skill lets the AI Assistant detect unusual patterns in your data directly from chat. Use this skill when you want to find unexpected spikes, drops, or outliers in a metric and review the results in a visualization. If data sharing is enabled, the assistant can also summarize the detected anomalies in natural language and suggest what to investigate next.

How It Works

The Anomaly Detection skill runs anomaly detection on your data and returns the result in the AI Assistant.

When activated, the assistant:

1. identifies the metric, time granularity, and other relevant parameters from the user request
2. uses the existing visualization flow to prepare the data
3. runs anomaly detection on the selected data
4. embeds the anomaly visualization in the chat
5. if data sharing is enabled, summarizes the detected anomalies and suggests what to investigate

The skill is designed for ad-hoc analysis in chat. It helps users investigate unusual behavior in their metrics without switching to a separate workflow.

Data Sharing

The Anomaly Detection skill always respects the AI Assistant data sharing setting.

When data sharing is disabled, the assistant can still run anomaly detection and display the visualization, but it does not use anomaly results for natural language interpretation.

When data sharing is enabled, the assistant can also describe the detected anomalies, summarize when and where they occurred, and suggest follow-up investigation steps.

Examples

Find Unusual Patterns Over Time
The user asks: Are there any unusual patterns in sales over time?
The AI Assistant then:

• identifies the relevant metric and time trend,
• runs anomaly detection,
• shows the anomaly visualization in the chat, and
• if data sharing is enabled, summarizes the unusual points in natural language.

Detect Revenue Anomalies
The user asks: Find anomalies in revenue by region.
The AI Assistant then:

• runs anomaly detection on the selected revenue trend,
• highlights unusual spikes or drops in the visualization, and
• if allowed by data sharing settings, explains what stands out and where to investigate.

Describe What Changed
The user asks: What looks abnormal in support ticket volume this month?
The AI Assistant then:

• runs anomaly detection on the metric,
• identifies unusual changes,
• displays the result in the chat, and
• if data sharing is enabled, summarizes the anomalies and suggests possible next steps.

Limitations

• This skill is intended for immediate, ad-hoc analysis
• It is not used for recurring monitoring or notifications
• Anomaly detection works best when the selected data contains enough history or variation
• Natural language interpretation of anomaly results is available only when data sharing is enabled

Error Handling

If anomaly detection cannot be completed, the assistant explains the issue and suggests a next step.

Examples include:

• insufficient data for anomaly detection
• no anomalies found in the selected data
• input that does not fit the anomaly detection requirements
• an error while running anomaly detection

In these cases, the assistant may suggest changing the metric, adjusting the analysis scope, or trying a different time range or granularity.