What's New — September 2, 2026

This release is focused on building deeper trust in your automated quality assurance and expanding your visibility across the entire contact center. Our AI now gets smarter before analysis even begins, automatically correcting transcripts to ensure fairer, more accurate scoring. With new integrations and more powerful validation tools, you can bring more of your operations into Ender Turing and be confident that the insights you see are the ones you can act on.

Executive Summary

Automatic Transcript Correction — Improve Auto QA accuracy with an AI that automatically finds and fixes misrecognized agent and company names in call greetings before they are scored.

Binotel Integration — Connect your Binotel cloud PBX directly from the Marketplace to automatically sync every recorded call for unified transcription, analytics, and quality assurance.

Highlights

Automatic Transcript Correction — Improve Auto QA Accuracy on Greetings

Speech recognition is powerful, but it can struggle with unique agent names or branded terms, leading to failed scorecard criteria and wasted time disputing results. When Auto QA scores a transcript with "Andrew" instead of "Andrzej," the agent is penalized for an error the system made.

▸ Improve scoring accuracy: Ender Turing now automatically corrects agent and company names in the call greeting before the transcript is analyzed. This removes a major source of false negatives from your Auto QA results.

▸ Reduce scorecard disputes: With corrected names appearing directly in the transcript, agents and team leads can trust that greeting and introduction checks are evaluated fairly, reducing manual reviews and appeals.

▸ Maintain data integrity: The system only replaces words that are a plausible phonetic match for a real name, and the original text is always preserved. Correction happens transparently and never delays a call from becoming available.

Before: Speech recognition would guess at names ("michal" instead of "Michał"), causing Auto QA to incorrectly fail the agent on their introduction.
After: The agent's name is automatically corrected against the user roster before Auto QA runs, so scores for greetings and introductions are accurate.
In practice: A QA Manager, Sarah, no longer has to manually overrule scorecards that failed an agent for their introduction, because the system now recognizes and corrects the agent's name automatically.

Get started: This feature is enabled automatically. Corrected names will appear in the transcripts of newly processed conversations.

Binotel Integration — Unify Your Call Center Analytics

Contact centers using Binotel have historically kept that call data siloed, making it impossible to apply consistent, platform-wide QA and analytics. Getting those calls into an analysis platform required manual exports or custom engineering work.

▸ Automate your data pipeline: Connect your Binotel account in a few clicks to have all recorded calls and their metadata automatically flow into Ender Turing, ready for analysis.

▸ Set up in minutes: The integration is self-service in the Marketplace. Simply enter your Binotel API key and secret to establish the connection, with no engineering support required.

▸ Get cleaner, more accurate data: The integration automatically identifies operators as agents, separates speaker channels for better transcription, and filters out unanswered or zero-duration calls to keep your dataset clean.

Before: Binotel calls were invisible to Ender Turing, requiring manual exports or custom scripts to analyze them alongside other channels.
After: All Binotel calls and recordings sync automatically, appearing in your conversation list for unified search, reporting, and QA.
In practice: A sales director, Mark, can now review performance across his entire team—including those on Binotel—using a single set of scorecards and dashboards in Ender Turing.

Get started: Navigate to the Marketplace in your settings, find the Binotel integration, and enter your API credentials to connect your account.

Improvements

▸ See the Conversations Behind AI Category Scores — When you evaluate an AI topic definition, the result counts are now clickable. Instantly drill down to see the exact list of conversations behind each metric (e.g., "incorrect matches") to understand precisely where your AI prompt is succeeding or failing, dramatically speeding up tuning and validation.

▸ More Reliable and Flexible Fireflies Integration — We've made our Fireflies integration more robust. You can now set a local timezone for your meetings so timestamps reflect your team's working hours. The sync is also more resilient, successfully importing meetings even if they are silent or have incomplete participant data from Fireflies.

▸ Faster Agent Metrics Dashboard Loading — The main dashboard's Agent Metrics table now loads immediately, even for organizations with very high call volumes. CSAT scores are now calculated on-demand only when you show or sort by a CSAT column, preventing them from delaying the entire page load.

▸ Consolidated Usage and Cost Reporting — All usage reporting is now centralized in a single "Usage" tab under System Settings. You can review your monthly license consumption (active agents, conversation counts) and your AI model costs side-by-side in one convenient, admin-only location.

▸ Consistent Filtering for "Not a Repeat Call" — When you filter for conversations that are not repeat calls within a specific timeframe, you now get consistent, accurate counts across the Conversations list, dashboards, and detailed reports. This ensures your first-contact resolution metrics and non-repeat call samples are trustworthy everywhere.