Fuel iX Copilots — Week 34 · August 17–23
A busy week with something for everyone — new AI models, smarter conversation handling, expanded file support, and sharper analytics tools, alongside a handful of reliability fixes that smooth out some rough edges.
-
Automatic context management for long conversations: Fuel iX now automatically summarizes conversation history when a session grows very long, so AI agents keep running smoothly during extended chats without any manual intervention needed.
-
Gemini 3.7 Flash and Gemini 3.5 Flash Lite now available: Two new Google Gemini models are available in the Copilot model picker under the Google group. Both support vision, tool calling, and a 1 million token context window.
-
YAML file upload support: You can now upload
.yamland.ymlfiles from your device, Google Drive, SharePoint, or directly in chat, making it easier to bring configuration files and structured data into your AI workflows. -
Improved usage chart granularity with click-to-zoom drill-down: Manager usage charts now automatically pick the best time scale (1-hour, 4-hour, or daily) based on your selected date range. Click any bar to zoom into hourly detail for that day. Charts also display in your local browser timezone for accurate day-by-day breakdowns, and your preferred chart type (bar or line) is remembered between sessions.
-
More reliable Google Drive folder disconnection: Disconnecting a Google Drive folder now correctly rolls back if something goes wrong during cleanup, preventing a broken state where the folder looked removed in the UI but its documents were still active in your Copilot. You'll also receive a clear error notification if the disconnect doesn't complete successfully.
-
Resolved orphaned records on user deletion: Deleting a user could previously leave behind leftover authentication records, causing inconsistent account states. The deletion process now confirms those records are fully removed and surfaces any failures clearly.
-
Improved AI code generation quality: Updated internal coding guidelines mean AI-assisted code generation now produces more consistent and maintainable output.
