This week we’ve landed improvements to the meeting recorder, suggested tasks, and added support for new state of the art models for larger queries.
Meeting recorder customization
You can now configure the Lightfield recorder’s appearance in meetings. It can display a logo image you upload, the default Lightfield logo, or disable the display entirely, which works well with “hiding non-video participants.”
Admins can now also enable or disable meeting recording with the Lightfield recorder for their entire workspace.

Easier to find suggested tasks
We now surface suggested tasks on related records to make them easier to track. Lightfield will also now capture more tasks from your meetings and emails.
We’re continuing to iterate on suggestion precision and coverage for tasks and for record updates, so keep the feedback coming.

New models for larger queries
We’ve added Gemini 3 Pro, Claude Opus 4.5, and GPT 5.1 as options when chatting with the workspace agent. Gemini 3 supports a ~1,000,000 token context window, and we have observed higher quality answers to queries of large data sets when using it. Note that it is currently slower with time to first token.


