What's New — October 2, 2025

This month's updates focus on automating your most time-consuming administrative tasks and delivering deeper, more reliable data across the entire platform. From zero-touch agent management to effortless QA framework setup and more intelligent keyword detection, you can now spend less time on manual configuration and more time improving contact center performance. These enhancements ensure your underlying data is always accurate and your analytics are built on a foundation you can trust.

Executive Summary

Automatic Agent Directory Sync — Keep your agent roster, team assignments, and active status perfectly in sync with your external workforce management system without any manual updates.

Import & Export QA Frameworks — Build, back up, and share entire scorecards and topic lists in seconds using simple Excel files, eliminating hours of manual setup.

Automatic Keyword Expansion for Tags — Capture every relevant conversation by letting the AI automatically generate all grammatical variations of your keywords, saving time and increasing accuracy.

Granular Detailed Scoring Exports — Analyze performance with greater precision by exporting scoring reports broken down by individual evaluation points, not just categories.

Reliable GenAI Prompt Editing — Confidently edit GenAI prompts in any mode, knowing your expert-level customizations are always preserved when you save.

New AI Provider: DeepSeek — Gain more flexibility in balancing cost and performance by using DeepSeek's powerful reasoning and chat models for your AI-driven analysis.

Highlights

Automatic Agent Directory Sync — Eliminate Manual Roster Management

Maintaining your agent roster in multiple systems is a tedious, error-prone task. When agents join, leave, or switch teams, your analytics platform can quickly fall out of sync, leading to inaccurate reporting and administrative headaches.

▸ Always-current agent data: Your agent list, team assignments, and active/inactive status are now automatically synchronized from your external directory. No more manual data entry.

▸ Accurate team-based analytics: With agent-to-team mappings always up to date, you can be confident that quality reviews and performance metrics are attributed to the correct team from day one.

▸ Effortless onboarding and offboarding: New agents appear in Ender Turing automatically when added to your directory, and departing agents are deactivated without admin intervention.

• Before: Admins had to manually add, deactivate, and reassign agents in Ender Turing whenever the roster changed in their primary workforce management system.
• After: Agent profiles, team assignments, and active status sync automatically from the external directory, requiring zero manual upkeep.
• In practice: A contact center manager moves an agent from the Sales team to the Support team in their directory; their analytics and quality scores are immediately attributed to the correct team in Ender Turing with no delay or manual changes.

Get started: This feature works automatically once your external directory integration is configured. Contact your administrator to enable it for your organization.

Import & Export QA Frameworks — Set Up Scorecards and Topics in Minutes

Building or replicating complex scorecards and topic lists has historically been a manual, click-by-click process. Transferring a QA framework to a new team or backing up your work required tedious re-creation or screenshots.

▸ Build and share with Excel: You can now export any scorecard or your entire topic list to an Excel file. This file can be shared, edited offline, or re-imported to another workspace.

▸ Drastically faster setup: Instead of creating scorecards and categories one by one in the UI, you can now define your entire QA structure in a spreadsheet and upload it in a single step.

▸ Create backups with confidence: Export your QA configurations to create a reliable backup before making major changes, knowing you can restore it at any time with a simple import.

• Before: Creating a new scorecard meant manually adding every category and point through the user interface, a process that could take hours.
• After: Define your complete scorecard structure in an Excel template and import it in seconds, or export an existing scorecard to replicate it instantly.
• In practice: A QA Manager needs to deploy a standard compliance scorecard to a newly onboarded team. They simply export the existing scorecard and import it into the new team's workspace, completing the setup in under a minute.

Get started: In the Scorecards or Topics section, use the new "Import from Excel" and "Export" options. You can download a template to get started with formatting.

Automatic Morphological Keyword Expansion — Find Every Mention, Effortlessly

To effectively track keywords with tags, you had to manually list every possible grammatical variation of a word—"refund," "refunds," "refunded," "refunding." This was time-consuming and often incomplete, causing you to miss relevant conversations.

▸ Automate keyword variations: Simply enter a base keyword, and Ender Turing's AI will automatically generate all its grammatical forms (plurals, tenses, declensions) for you.

▸ Increase tag accuracy: By covering all morphological cases, you ensure that no relevant conversation slips through the cracks due to an unlisted word form.

▸ Save significant setup time: Eliminate the manual effort of brainstorming and entering dozens of word forms, especially for highly inflected languages.

• Before: Creating a tag to find discussions about "cancellations" required you to manually enter "cancel," "cancels," "cancelled," "cancelling," and "cancellation."
• After: You enter "cancel" and "cancellation," and the system automatically finds all related word forms in conversations.
• In practice: A QA Manager creates a tag for "complaints." The system automatically expands it to catch conversations where customers say "I complained," "we are complaining," or "I have a complaint," without the manager having to enter each phrase.

Get started: This feature is on by default. When you create or edit a tag, the expansion happens automatically in the background. Admins can manage this feature in Settings.

Granular Detailed Scoring Exports — Dig Deeper into Evaluation Results

Previously, detailed scoring exports grouped results by category, making it difficult to analyze performance on specific evaluation criteria or easily locate comments tied to individual points.

▸ Analyze by scoring point: The detailed scoring export now offers a view broken down by individual scoring points, giving you the most granular look at agent performance.

▸ Access structured comments: Each category and point now gets its own dedicated comments column in the export, so you can find specific feedback without scanning a single, merged text block.

▸ Enjoy a faster UI: Detailed data is now reserved for the export file, which speeds up the loading time of scoring reports within the application.

• Before: The scoring export grouped scores by category, and all reviewer comments were merged into a single column.
• After: You can choose an export that breaks down scores by individual evaluation point, with reviewer comments neatly organized into separate columns for each point.
• In practice: A QA Manager exports a point-level report to identify that while the "Professionalism" category score is high, agents consistently lose points on the specific "Used approved greeting" criterion.

Get started: When you download the "Detailed Scoring XLS" report from the agent scoring section, the enriched data appears automatically in your downloaded file.

Reliable GenAI Prompt Editing — Save Edits in Any Mode Without Losing Work

Editing complex GenAI prompts was risky. A critical bug could cause any expert-level customizations to be silently deleted if you made a change in the simpler "Standard" editing mode, forcing you to use workarounds or risk losing valuable work.

▸ Preserve your expert content: You can now safely make changes in Standard mode. The system will only update the standard sections of your prompt, leaving all expert-only content untouched.

▸ Save edits reliably: The "Update" button now correctly activates whenever you make a change in Standard or Expert mode, ensuring you can save your work without friction.

▸ Edit with confidence: Freely switch between editing modes while crafting prompts, knowing that all your configurations are protected and will be saved reliably.

• Before: Saving a prompt in Standard Mode would overwrite any custom logic you had added in Expert Mode, silently deleting your work.
• After: Saving in Standard Mode now correctly preserves all expert-only content, allowing you to use both modes safely.
• In practice: A QA manager fine-tunes a prompt with expert-level instructions, then later switches to Standard mode to adjust a basic setting. After clicking "Update," their expert customizations are now safely retained.

Get started: This fix is applied automatically. Continue editing your prompts in Analytics > GenAI Studio as usual.

New AI Provider: DeepSeek — More Flexibility for Your AI Workflows

Your AI needs for quality, cost, and performance are unique. We are expanding your options by integrating another leading AI provider directly into the platform.

▸ Choose the best model for the job: You can now select DeepSeek's powerful reasoning and chat models to power Auto QA, conversation summaries, categorization, and other AI features.

▸ Optimize for cost and performance: Add another high-quality AI provider to your toolkit, giving you more flexibility to choose the model that best fits your requirements for any given task.

▸ Connect to custom endpoints: For organizations using self-hosted or proxy gateways, you can configure a custom API base URL for all DeepSeek requests.

• Before: AI provider options were limited to OpenAI, Gemini, Azure, and Anthropic.
• After: DeepSeek is now available as a first-class provider, selectable for any prompt or AI-driven workflow.
• In practice: An administrator configures a new Auto QA prompt to use a DeepSeek model, choosing it for its balance of high performance and cost-effectiveness for that specific evaluation task.

Get started: An administrator can go to Analytics Configuration settings to enter a DeepSeek API key. Once configured, DeepSeek will appear as a provider option when editing any prompt.

Improvements

▸ Audit Logging for All User Actions — For enhanced security and compliance, all data-changing actions (like creating users, updating scorecards, or changing settings) are now automatically logged. This provides a complete audit trail of who made what change and when, with sensitive data automatically redacted.

▸ Improved Genesys Data Integrity — We've resolved issues that could cause some Genesys Cloud conversations to be missed during synchronization. Transferred calls are now also correctly attributed to the last handling agent, not the first, ensuring your quality and performance data is accurate.

▸ Weighted Quality Scores on Main Dashboard — The Main Dashboard table now includes columns for both "Avg score (weighted)" and "Avg score (simple)," giving you a complete quality overview at a glance. You can also customize which columns are visible to focus on the metrics that matter most.

▸ Enhanced Analytics Charting with "Leads" Metric — You can now track unique customer contacts ("Leads") as a metric in your analytics charts, alongside sessions and cost. Each dataset in a chart can also be configured with its own metric and chart type (bar or line), allowing for richer data comparisons.

▸ Faster, More Flexible Executive Dashboards — C-Level Boards are now more responsive, with faster loading when you switch between boards. You can also add new "Volume Meter" widgets to track key numbers like conversation counts and spend, and find charts more easily with a new search field.

▸ Faster, Smarter, and More Cost-Effective AI Processing — AI-powered features like Auto QA and Summarization now run more efficiently. By intelligently caching requests and parallelizing tasks, we've made analysis faster and reduced redundant processing, saving you time and lowering operational costs.

▸ Correct Team Assignments and Data Visibility — Agents are now automatically assigned to the correct team based on their group, ensuring data is accurately associated across the platform. This also resolves an issue where team leads could not view quiz results for agents on their own team.

▸ ViciDial Integration Supports Alphanumeric Agent IDs — Call recordings from ViciDial systems that use letter-based or mixed agent IDs (e.g., `VUAVKI15`) are now correctly imported and processed, ensuring complete data capture for all agents.

▸ Smarter Anonymization for Company Names — Our anonymization engine is now more precise, preserving brand and company names that contain numbers (e.g., "Credit365") in transcripts while continuing to mask sensitive, standalone number sequences.

▸ Streamlined Automation Rule Management — The version history for automation rules now updates immediately after saving, so you don't have to reload the page. Additionally, you can now use "Select All" and "Deselect All" buttons to bulk-manage topic categories when configuring actions.

▸ Improved Agent Self-Review Experience — The agent view is now cleaner and more focused. Agents can see conversation topics directly on each call and the scorecard selector now only displays scorecards relevant to their team, reducing clutter.

▸ Clickable Metadata Links in Conversation View — More metadata from your integrated systems, such as order or loan URLs, are now automatically detected and displayed as clickable links in the conversation detail panel, speeding up investigations.

Fixes

▸ Improved reliability for conversation processing ensures that interactions that fail temporarily are automatically retried.

▸ Drag-and-drop reordering of points within scorecard categories is now fully functional and saves reliably.

▸ TODO reminder emails are now correctly consolidated into a single summary email instead of multiple fragmented messages.

▸ The Agent Dashboard now correctly displays a quality score of "0" instead of incorrectly showing it as "N/A".

▸ The agent filter now correctly finds sessions that have no agent assigned when you invert a selection of all agents.

▸ An error that could occur in EnderChat (formerly AI Chat) when using prompts that did not have a description has been resolved.

▸ The EnderChat interface is now smoother, with a sticky header in the conversation list and more reliable message ordering.

▸ The built-in AI assistant has been rebranded from "ChatGPT" to "EnderChat" for a more consistent brand experience.