Ender Turing Updates — 2026-10-02 to 2026-10-15

Important changes

EnderGPT answers questions using PDF attachments

Category: feature
Importance: high
What changed: EnderGPT reads attached PDFs alongside conversation content, identifies source filenames and conversations, and explains unreadable or partially reviewed documents.
Customer benefit: Investigate contracts, invoices and other attached evidence without reading each file manually.
Highlights:

• Refer to a conversation by name, ID or link; the Has attachments filter narrows conversation lists, reports and EnderGPT selections.

Clearer conversation headers and one-click number copying

Category: improvement
Importance: high
What changed: Conversation headers support copying customer numbers behind partner names, clearer action menus, separate expandable Topic and CRM status rows, and easier conversation navigation.
Customer benefit: Find key details and copy the right number with less effort.
Highlights:

• Copy confirmations appear only after success, contrast is clearer in both themes, and a sidebar sun or moon control switches themes.

Choose visible tools in Settings and Features

Category: feature
Importance: high
What changed: Administrators manage optional tools from one Features page with descriptions, contextual help, and Save and Cancel controls.
Customer benefit: Keep navigation focused on the tools your team uses while retaining existing data and configurations.
Highlights:

• Disabling Appeals blocks new appeals while existing appeals remain available for review and resolution.

Deactivated agents stay inactive after edits

Category: fix
Importance: medium
What changed: Editing an agent's name, phone or team no longer reactivates the agent or linked user, including after bulk deactivation with an editor open.
Customer benefit: Keep access disabled until an administrator deliberately restores it.

Summary-triggered Enders avoid duplicate actions

Category: fix
Importance: medium
What changed: Enders using New summary added with Search and Summary prompt check the triggering summary and avoid repeated comments or pending notifications for the same conversation and prompt.
Customer benefit: Receive one relevant action instead of repeated alerts or matches from unrelated content.
Highlights:

• The editor explains this filter combination and rejects unsupported search terms; ordinary Search and standalone Summary prompt filters keep their behavior.

Other changes

Read the latest answer from its beginning

Category: improvement
Importance: medium
What changed: Finished EnderGPT chats and scheduled reports open at the start of the latest answer, with controls to jump to the beginning or end of the chat.
Customer benefit: Read long reports without scrolling back from the bottom or losing your position during generation.
Highlights:

• Controls stay clear of the text on wide and narrow screens, including for read-only recipients.

Deactivate accounts before sign-in setup is complete

Category: fix
Importance: medium
What changed: Administrators can deactivate agents and users with pending invitations, and later invitation acceptance or agent-link changes do not restore access.
Customer benefit: Offboard people immediately and keep accounts disabled until deliberately reactivated.
Highlights:

• Repeating an activation or deactivation already in effect succeeds without an unnecessary error.

Download individual agent scoring summaries

Category: improvement
Importance: medium
What changed: Managers can export an agent's scoring table as PNG, XLSX, CSV or TXT, with detailed Excel reports by score category or point.
Customer benefit: Share scoring results that match the selected agent, scorecard, mode, filters and expanded category.
Highlights:

• Detailed exports include comments and conversation details; agents no longer receive an unrelated conversation-opening error, and transcript correction actions follow permissions.

More reliable single sign-on checks and permissions

Category: fix
Importance: medium
What changed: Brief identity-service delays no longer unnecessarily interrupt pages and notifications, and user-management permission changes preserve the user's other roles.
Customer benefit: Stay productive through temporary connection problems with correctly applied access changes.

Conversation score counts include automated reviews

Category: fix
Importance: medium
What changed: Conversation lists, Info panels and agent Scored Records show the number of actual manual and AutoQA evaluations, counting each submitted automated scorecard separately.
Customer benefit: See how many reviews contribute to a conversation's score without counting unsubmitted defaults.

Clearer validation for incomplete AI model settings

Category: fix
Importance: medium
What changed: Partially saved tasks show Default and save correctly; incomplete Azure OpenAI details are marked unconfigured and required fields are flagged immediately.
Customer benefit: Correct incomplete settings before AI features try to use them.
Highlights:

• Saving clears leftover values, and Ender Chat and transcript correction skip incomplete setups.

Customer charts include single-conversation customers

Category: fix
Importance: medium
What changed: Customer chart cells count existing conversations correctly, including a single conversation, and drill-downs and Excel exports match the displayed values.
Customer benefit: Find one-off customers, including costly cases, without false zero counts.
Highlights:

• The default top-20 view still prioritizes repeat customers.

Recommended AI models update automatically

Category: improvement
Importance: medium
What changed: AI features use the latest Ender Turing recommended models by default, while an organization's explicit model choice takes priority.
Customer benefit: Keep AI tasks on maintained defaults without managing a model for every prompt.
Highlights:

• Prompts follow their feature's model; Azure details are trimmed and validated, and custom deployments remain selectable.

Faster cloud security maintenance

Category: fix
Importance: medium
What changed: The cloud platform checks for operating-system security updates daily and applies available updates through rolling service replacement.
Customer benefit: Receive security maintenance promptly while services remain available.

Accurate totals across customer chart datasets

Category: fix
Importance: medium
What changed: Charts grouped by customer show complete values in every dataset, with matching Excel exports and clear errors if a report cannot be completed accurately.
Customer benefit: Compare customer values and calculate ratios without false zero totals.

Steadier automation and categorization under load

Category: fix
Importance: medium
What changed: Large automation runs complete more reliably, partner filters run faster, and follow-up analysis starts only after its triggering action is saved.
Customer benefit: Keep categorization, AutoQA and summaries current with fewer peak-volume delays or repeated jobs.
Highlights:

• Reverted actions do not start their follow-up work.

Show why an AI score passed

Category: feature
Importance: medium
What changed: AutoQA can retain positive explanations as well as reasons for failed or not-applicable criteria, with a switch to show positive explanations.
Customer benefit: Understand successful agent behavior while keeping review screens uncluttered by default.
Highlights:

• Existing reports, exports, editing and appeals retain their behavior unless the feature is enabled.

Total LLM spend on the Usage page

Category: improvement
Importance: medium
What changed: The LLM Usage heading displays the total USD cost for the current dates, prompts, active or archived view and grouping, even when the table shows tokens.
Customer benefit: See total spend without adding rows manually, including historical usage from removed apps or unknown prompts and models.

Complete notification emails through corporate mail servers

Category: fix
Importance: medium
What changed: Notification emails retain their full plain-text and HTML message content when processed by strict corporate mail systems.
Customer benefit: Read assignments, mentions and alerts directly in your inbox instead of receiving an empty message.

Previous conversation interval works in more reports

Category: fix
Importance: medium
What changed: The Previous conversation interval filter now works in tag, reviewer performance, SLA and Topic reports and widgets that previously returned errors.
Customer benefit: Extend repeat-contact analysis across reports while preserving results in reports that already supported it.