View a markdown version of this page

Configuring Insights - Amazon Connect Decisions

Configuring Insights

Insights configuration is how you teach Amazon Connect Decisions what to watch for, how to interpret what it finds, and how to prioritize what surfaces to your team. Configuration applies to both demand and supply monitoring. There are four areas to work through in sequence: Knowledge Sources, Detection (Metrics and Rules), Root Causes and Recommendations (Guidelines), and Prioritization (Severity). A preview is available at each step so you can validate before activating any configurations.

To open configuration: From the home page, navigate to Insights in the left-hand navigation, then select Configuration.

The four configuration areas

  • Knowledge Sources: Share SOPs and documented business best practices so Amazon Connect Decisions understands your operational context before generating metrics, rules, and guidelines.

  • Detection: Define the metrics and rules that determine when an insight is triggered for demand or supply monitoring.

  • Guidelines: Shape how root cause analysis and recommendations are generated so they align with your business practices.

  • Severity: Assign financial impact to insight types so Amazon Connect Decisions can rank Insights by business priority.

How configuration connects to Insights output

Each area feeds the Insights experience directly. Detection determines what surfaces; guidelines shape the analysis and recommended actions; severity determines the order in which Insights appear. Changes to any area take effect as soon as the relevant items are set to Ready status.

Best practices

  • Start with clear monitoring objectives: Before creating metrics and rules, define what supply chain performance issues matter most to your operations. This helps you configure focused, actionable monitoring rather than generating excessive alerts.

  • Leverage your onboarding teammate: Use natural language to describe monitoring needs rather than attempting to write technical specifications. Ask questions like “Can we make this more sensitive to A-class products?” or “What does this field do?” Connect Decisions agents translate your business requirements into appropriate configurations.

  • Preview extensively: Always preview metric and rule changes before activating them. Use the preview function to:

    • Validate that metrics calculate correctly against your data

    • Confirm that rules trigger insights at appropriate thresholds

    • Filter preview results by products and sites to understand impact across your supply chain

    • Iterate on configurations until preview results match your operational needs

  • Use explainability instead of SQL: When reviewing how metrics and rules operate, use the Explainability tab rather than examining SQL queries directly. This helps you understand the logic in plain language without technical complexity. When you see something that doesn’t match your expectation work with the agent to correct that via explainability.

  • Iterate with draft status: Take advantage of the Draft status to configure and test metrics and rules without impacting production insights. Only change status to Ready when you’re confident in the configuration.

  • Balance sensitivity: Set thresholds that catch meaningful insights without generating excessive noise. Use preview to find the right balance between alert coverage and operational manageability.

  • Start simple, then expand: Begin with a few key metrics and rules for your highest-priority monitoring needs. Once you’re comfortable with the configuration workflow, expand to additional monitoring areas.

  • Review and refine regularly: After activating metrics and rules, monitor the quality and relevance of generated insights on your Planning Intelligence home page. Return to Monitoring Configuration to refine settings based on operational experience.