Ender Turing Updates — 2026-05-16 to 2026-06-01

Important changes

AI-assisted tuning for Categories

Category: feature
Importance: high
What changed: A new workflow uses AI to help you create, tune, and evaluate category definitions based on labeled example conversations.
Customer benefit: This significantly speeds up the setup of new categories and continuously improves their accuracy with AI-suggested refinements.
Highlights:

• Generate an initial definition from a small set of labeled examples.
• Review and approve AI-suggested improvements in a guided queue.
• Track the precision, recall, and accuracy for each version of a category's definition.

Flexible visibility controls for EnderGPT Chats

Category: feature
Importance: high
What changed: Administrators can now assign a new permission to specific roles, granting users visibility into all EnderGPT chats across the organization.
Customer benefit: This gives you granular control over who can view chat history, replacing previous all-or-nothing access rules.
Highlights:

• The new permission is available in the Role editor under the 'EnderGPT Chats' section.
• Users with the Manager role can now create their own EnderGPT chats by default.

Topic accuracy and version history improvements

Category: improvement
Importance: medium
What changed: The Topics list now displays the accuracy score for each topic, generates definitions in your system's language, and shows which user made manual edits in the version history.
Customer benefit: This allows you to quickly prioritize which topics to refine, understand definition changes at a glance, and reduce manual translation efforts.
Highlights:

• Accuracy scores are now visible directly on the main topics list.
• Manual edits in the version history now show the name of the user who made the change.
• A streamlined definition format results in faster and more cost-efficient topic evaluation.

Fixed missing repeat calls in analytics and filters

Category: fix
Importance: medium
What changed: Repeat calls received during a specific early morning time window are now correctly identified and included in analytics and filters.
Customer benefit: This ensures your repetitive-call reports and filters are complete and accurate, reflecting the true volume of repeat contacts at all hours.

Smarter notifications for empty Scheduled Chat runs

Category: improvement
Importance: medium
What changed: Scheduled Chats now send a dedicated notification instead of an AI-generated summary when the filter finds no matching conversations.
Customer benefit: This avoids confusing 'no data' summaries, reduces unnecessary AI costs, and provides direct links to review the filter or edit the schedule.
Highlights:

• The new notification is titled 'No conversations matched' to distinguish it from successful runs.
• Notifications include one-click links to 'Review filter' and 'Edit schedule'.
• Empty runs no longer create a session in your chat history.

Other changes

Source links are preserved when copying chat messages

Category: fix
Importance: medium
What changed: When you copy an assistant message from Chat, links to conversations and agents are now included and remain clickable when pasted into other applications.
Customer benefit: This allows you to share findings with full context, enabling recipients to jump directly to the cited conversations or agent profiles.
Highlights:

• Links work when pasting into rich-text applications like email, Slack, or Google Docs.

More reliable and accurate Ender Chat answers

Category: improvement
Importance: medium
What changed: Ender Chat responses are now more trustworthy, with improved validation to prevent fabricated links and ensure cited quotes match the original transcripts.
Customer benefit: This increases confidence in AI-generated summaries and reports, reducing the need to manually verify every cited source.
Highlights:

• The chat assistant will now only generate links to conversations, agents, or filtered views within the Ender Turing platform.
• Quotes are now validated against source transcripts to ensure accuracy.

Accurate agent and queue data for Genesys integrations

Category: fix
Importance: medium
What changed: For customers with Genesys integrations, conversation metadata such as agent and queue names is now correctly isolated in multi-organization environments.
Customer benefit: This resolves a data-mixing issue and ensures conversation details, reports, and Auto QA scores are always attributed to the correct organization.

More reliable Genesys call recording downloads

Category: fix
Importance: medium
What changed: The Genesys integration now automatically refreshes expired download links for call recordings, preventing download failures during processing delays.
Customer benefit: This ensures all your Genesys call recordings are reliably ingested for analysis, even during high-volume periods.

Scorecard category totals now correctly exclude critical points

Category: fix
Importance: medium
What changed: In the scoring panel, category point totals now correctly align with the overall session total when your scorecard is configured to exclude critical points.
Customer benefit: This ensures scoring calculations are consistent and easy to understand, removing confusion when reviewing an agent's evaluation.

New UI for single sign-on user provisioning

Category: feature
Importance: medium
What changed: You can now enable automatic user account creation for single sign-on (SSO) directly from a new toggle in the Authentication settings.
Customer benefit: This allows new users from your identity provider to be provisioned on their first sign-in without manual account creation by an administrator.
Highlights:

• New users are automatically created and assigned the Agent role by default.
• A new 'OIDC linked' badge on the user management screen clearly identifies accounts managed via SSO.