What's New — June 2, 2026

What's New — June 2, 2026

This release marks a significant step toward a more autonomous quality program. Instead of manually tuning analytics and restricting powerful tools, you can now leverage AI to automatically refine conversation topics and securely empower your entire management team to get insights from EnderGPT Chat.

Executive Summary

AI-Powered Topic Auto-Tuning — Automatically generate, calibrate, and improve the accuracy of your conversation topics with a new AI-assisted workflow that turns days of work into minutes.

Flexible Permissions for EnderGPT Chat — Empower managers to run their own AI queries and grant specific users organization-wide visibility into all chat activity through a new, flexible role permission.

Highlights

AI-Powered Topic Auto-Tuning — Build Accurate Analytics in Minutes, Not Days

Building and maintaining accurate conversation topics has always been a manual, time-consuming process of trial-and-error, with no clear way to measure or improve performance.

▸ Generate definitions instantly: Create a starting topic definition from just a handful of example conversations instead of writing complex rules from scratch.

▸ Let AI do the tuning: The system analyzes misclassifications and automatically suggests improvements to your topic definitions for you to review and approve.

▸ Track performance over time: See accuracy, precision, and recall metrics for every version of a topic, and view accuracy at a glance right from the main topics list.

• Before: You manually wrote topic rules, guessed at their accuracy, and discovered mistakes only through random spot-checks.
• After: You generate a baseline topic from examples, approve AI-suggested refinements from a queue, and track measured accuracy on a dashboard.
• In practice: A QA manager, Sarah, launches a new "Customer Retention Risk" topic by labeling 15 calls. She accepts two AI-suggested definition tweaks, bringing accuracy to 95% before her lunch break.

Get started by navigating to the Topics section. Create or edit a topic to access the new LLM Definition workflows, review queue, and evaluation history.

Flexible EnderGPT Chat Permissions — Empower Managers and Control Visibility

Access to EnderGPT Chat was often restrictive, preventing managers from running their own analyses and making it difficult for administrators to grant oversight capabilities to specific users.

▸ Empower your managers: Team Leads and Managers can now create their own EnderGPT chats by default, allowing them to investigate team performance and conversation trends directly.

▸ Grant organization-wide visibility: A new role permission lets you give specific users (like compliance officers or senior admins) the ability to view all EnderGPT chats across the company.

▸ Simplify access control: Manage all chat permissions from a clearly labeled "EnderGPT Chats" section in the Roles editor, removing the need for hidden rules or support requests.

• Before: Managers could only view chats shared with them, and cross-user visibility was limited by inflexible, hidden rules.
• After: Managers can create their own chats, and admins can assign a "View All Chats" permission to any user through a custom role.
• In practice: Director John wants his two lead analysts to audit all AI chat usage. He creates a new "Chat Auditor" role with the new permission and assigns it to them in under a minute.

Get started by going to Settings → Roles to find the "EnderGPT Chats" section and configure permissions. Managers will automatically have creator access with no action needed.

Improvements

▸ Smarter, More Reliable Ender Chat Answers — Ender Chat now produces more trustworthy answers. The system validates that all quotes are from the actual transcript and ensures all source links point to real conversations within Ender Turing. Plus, when you copy a chat message, all source links are preserved, making it easy to share findings with full context in emails or documents.

▸ Simplified SSO User Provisioning — Administrators can now enable automatic user account creation for single sign-on (SSO) directly from the Authentication settings page. New users signing in via your corporate identity provider (like Microsoft Entra ID) will have accounts created on-the-fly, eliminating manual setup and streamlining onboarding.

▸ Smarter Alerts for Empty Scheduled Reports — Scheduled Chat reports that find no matching conversations will no longer create a confusing "no data" summary. Instead, you'll receive a clear notification that the filter returned no results, with direct links to review the filter or edit the schedule, saving time and avoiding unnecessary AI processing.

▸ More Reliable Genesys Integration — Your Genesys conversation data is now more complete and secure. We’ve resolved an issue to ensure call recordings are downloaded successfully even during high-volume periods, and fixed a caching error in multi-organization setups to guarantee agent and queue data is always attributed to the correct tenant.

▸ Accurate Repeat Call Detection — We've fixed a bug that caused some repeat calls during a specific morning window to be missed by our analytics. The "Interval since previous call" filter and repetitive-call reports are now accurate 24/7, giving you a complete picture of recurring customer issues.

Fixes

▸ Corrected an issue where category totals in scorecards did not respect the "Exclude Critical Points" setting, ensuring scoring numbers are now consistent across the entire review panel.