Meet Ender, the AI Quality Assurance Specialist who will join your QA team and help it shine!
.svg){kind=small}
This month's update is about connecting your insights to the outside world and digging deeper with more reliable AI. We've overhauled our automation and chat analytics to give you the power to send precisely formatted data to any system and to trust that the answers you get are comprehensive and accurate. These enhancements bridge the gap between analysis and action, helping you embed quality insights directly into your contact center's daily operations.
Customizable Webhook Payloads — Shape automation data to match any external system's format using a powerful new in-platform transformation engine, eliminating the need for middleware.
A Reimagined EnderGPT Chat Experience — Get deeper, more accurate insights with a completely redesigned chat UI and a fixed analysis engine that now processes 100% of your filtered conversations, not just a small sample.
Enhanced Chat Privacy — Your private chat sessions are now truly confidential and are no longer visible to administrators, ensuring a secure space for analysis and exploration.
Previously, sending data from Ender Turing to other tools required you to accept our fixed data format, often forcing you to build an intermediate service just to translate the payload.
▸ Shape data on the fly: Use the powerful JSONata expression language to rename fields, restructure objects, and transform values right inside the webhook action. Send data in the exact format your target system requires.
▸ Test and validate before you save: A new preview and test function lets you see the final payload against real conversation data and even send a live test to your endpoint to verify the response. Catch errors during configuration, not in production.
▸ Eliminate middleware: Integrate directly with CRMs, ticketing systems like Jira, business intelligence tools, or even simple chat notifications without writing a single line of "glue code."
Get started by navigating to any automation, adding a Webhook action, and selecting the "Custom Payload" option.
EnderGPT Chat is a powerful tool for asking complex questions of your conversation data, but its analysis could be incomplete and the interface didn't always keep you informed.
▸ Get the full picture: EnderGPT now correctly analyzes all conversations that match your filters, especially when filtering by CSAT. A critical fix ensures you get comprehensive insights from your entire dataset, not just a small, misleading sample.
▸ Know what's happening: A redesigned interface provides clear progress updates during analysis, cycling through stages like "Analyzing conversations..." and "Compiling results..." so you're never left wondering if it's still working.
▸ Start faster with a better UI: The chat composer is now more inviting, with example questions to guide you. The entire experience is more polished, with better theme support and clearer display of active filters.
Get started by navigating to EnderGPT Chat from the main menu. The improved experience is now active for all users.
For chat to be a useful tool for exploration, you need to trust that your private analyses remain private. Previously, administrator-level users had visibility into all chat sessions across the organization.
▸ Ensure confidentiality: Private chats you create are now visible only to you. They will not appear in anyone else's list, including administrators, unless you explicitly share them.
▸ Maintain consistent rules: Administrators now follow the same visibility rules as every other user—they see their own chats and any chats that have been explicitly shared with them.
This change is now active for all users. No action is required.
▸ Faster, Smarter, and More Accurate Dashboards — Dashboards now load significantly faster on repeat visits thanks to smarter caching, and C-Level Board widgets no longer flicker on refresh. We've also improved the accuracy of Interaction Intensity scoring to better reflect agent pressure, and fixed navigation so managers always see the Dashboard menu and C-level boards are fully scrollable.
▸ Self-Service Exports and Clearer Review Context — Managers and team leads with the right permissions can now export scorecards and topic lists directly as XLS files. Scorecards now display the review date in the footer for quick reference, and all exported reports are cleaner, with correct formula calculations, accurate review dates, and timestamped comments.
▸ More Reliable AI-Powered Scoring and Categorization — We've made our AI engine more robust. AutoQA now uses a centrally managed prompt for more consistent results, while automatic CSAT, Intensity, and Topic categorization now feature stricter validation and better error handling to ensure scores and categories are applied reliably, even under high load.
▸ Cleaner, More Intuitive Filtering — Dashboard filter selections now stay on a single, tidy line, even with long names. More importantly, any filter option you've already selected now conveniently floats to the top of the dropdown list, eliminating the need to scroll and search for your active filters.
▸ Safer Prompt and Ender Management — To prevent accidental disruptions to your automations, you'll now see a confirmation dialog when archiving a Summary Prompt. This dialog lists all the active Enders that rely on that prompt, allowing you to confirm the change or cancel to keep your automations running.
▸ Webhook automations now stop firing immediately when disabled and deliver more reliably under load.
▸ Single Sign-On (SSO) is more reliable, tolerating minor clock differences between systems and providing a clear error message on unauthorized attempts instead of a frozen screen.
▸ Improved data privacy for AI features, preventing sensitive conversation content from appearing in error logs.
▸ Existing call transcripts are now protected from being accidentally overwritten with empty data during processing retries.
▸ Conversation processing is now faster by default for most customers by intelligently skipping an unnecessary audio analysis step.
▸ Enhanced platform security by updating a core networking dependency to better protect requests.
The AI terminology chaos is real. Your "divide and conquer" framework is the clarity we needed.
Finally, a clear way to cut through the AI hype. It's not about the name, but the problem it solves.
