2026.03 Conversation Insights Agent
Release Date: 13 April 2026
The Conversation Insights Agent 2026.03 Release introduces the following features.
New Features
Introducing Conditional Logic in Scorecards
Conditional Logic helps auto evaluate Agents' and Teams' performance by applying multiple conditions and rules to the Scorecard. It is configured under Scorecard > Question. When enabled, you can have up to five conditions, each containing up to five rules, with auto evaluation as the First rule in each condition. Each rule can be based on a Fact or Metadata, with only Boolean Facts and Call Duration Metadata currently supported. Conditions and rules are evaluated in the order they appear. If no conditions are matched during evaluation, the Default Condition's score is applied as a fallback.
The Quality Analysts can view the applied conditions in Evaluations > Scorecard and modify them when Assisted Manual QM is enabled for the Scorecard. Also, the Conversations details show the applied conditions.
Note
Conditional Logic requires a Fact to be selected in Auto Evaluation and is disabled by default.
For more information, refer to Create Question and Manual Evaluation.
Conditional Logic in Scorecard-Questions
![]() |
Conditional Scorecard in Evaluations
![]() |
Enhancements
Conversation Insights Agent is now available in GCP
Conversation Insights Agent is now expanded to GCP (Google Cloud Platform) in the Middle East (Saudi Arabia). This release includes AWS-to-GCP infrastructure migrations (S3 → GCS, ElastiCache → Memorystore, RDS → Cloud SQL) and GitOps-managed service deployments via Flux, along with activities to resolve NFS/Filestore migration blockers and cluster-mode Redis provisioning.
Introducing a View to Identify and Measure Repeat Callers
In Custom Analytics, a new view, Repeat Callers Metadata, is introduced to support Recontact Analytics reporting based on Customer Id. When the same customer calls more than once to resolve the same issue/topic, they are considered repeat callers. It helps to identify and measure repeat callers within a configurable time period, segmented by topic, agent, and time distribution.

Constraints and Limitations
There are no constraints and limitations specific to this release.
Bug Fixes
The following issues were fixed in this release:
Item | Description |
|---|---|
1 | When users opened a conversation page via a direct conversation ID link from Custom Analytics (dashboards), audio playback did not work. This issue is now fixed. |
2 | In Custom Analytics, dashboards showed zero conversation counts when users applied the is on or after date filter to the Total Conversation Count metric. This issue has been fixed. |
3 | On the call details page, the Customer Talk duration metric was showing negative values. This issue has been resolved. |
4 | On the call details page, conversation turns containing scoring evidence were not displayed, leaving users unable to locate or navigate to the part of the call where a rating was given. This issue has been resolved. |

