- Newest
- Most votes
- Most comments
While the initial response correctly identifies S3 as the destination for individual evaluation data, there are two critical practical points to consider for a complete solution:
1. Data Format and Accessibility
The evaluations are exported to S3 as individual JSON files. If you are looking for a consolidated list (like a spreadsheet) for analysis, you won't be able to simply "download" one file. Each evaluation session creates its own object in the bucket.
2. Transformation via Amazon Athena
To effectively export and analyze these scores across all agents, the best practice is to use Amazon Athena.
- Step 1: Point Athena to your Contact Lens evaluation S3 bucket.
- Step 2: Define the schema for the JSON output (Amazon Connect provides standard templates for this).
- Step 3: Run a SQL query to join the individual scores into a single table.
- Step 4: Export the query result as a CSV file for use in Excel or BI tools like QuickSight.
Summary of the Evaluation Event Stream:
If you need this data in real-time, ensure that Contact Evaluation Updates are enabled in your Amazon Connect instance settings. This triggers the event flow to EventBridge and the subsequent delivery to S3 via Kinesis Data Firehose (if configured) or the direct S3 export.
Yes, you can export individual quality evaluation scores in Amazon Connect Contact Lens. When you enable contact evaluations in the Amazon Connect console, you'll be prompted to create or choose an S3 bucket where the individual evaluation data will be stored.
Each evaluation is exported as a separate file to your S3 bucket, containing detailed information including the individual scores for each section and question within the evaluation form. The export includes metadata such as the evaluation ID, evaluator information, timestamps, and the overall score, as well as granular details like section scores, question-level scores, answers, and question text.
To access this detailed data, you'll need to retrieve the files from the S3 bucket you configured. These exported evaluations contain the individual scores you're looking for, rather than just the aggregated averages shown in the dashboard reports.
Additionally, if you need to view evaluation details through the CloudWatch logs, you can fetch evaluation events that include per-session, per-evaluator scores with specific score values and labels for each evaluation performed.
This approach allows you to see the complete evaluation history for an agent with all individual scores, rather than just the averaged metrics displayed in the standard reports.
Sources
Agent evaluation form output in Amazon Connect - Amazon Connect
Batch dataset runner - Amazon Bedrock AgentCore
Relevant content
- AWS OFFICIALUpdated 2 years ago
