Schedule campaigns and compare product bets
Schedule reviewed email campaigns, weigh equally ranked Decision Lab bets, and inspect more complete product context and source details.
What changed
- Schedule a reviewed email campaign for a chosen date, time and one named time zone. Change or remove the schedule before dispatch.
- Review equally ranked Decision Lab bets, their reasons and evidence, then choose a follow-up for yourself to review and send.
- Inspect clearer source identity and quote context in Discovery, and use fuller scoped Context reads in Workspace Chat.
Schedule a reviewed email campaign
Admin can schedule a reviewed email campaign batch for a chosen date, time and one named time zone. Change or remove the schedule before dispatch. A worker checks for due sends once a minute; actual inbox arrival can take longer.
Review the Decision Lab slate
Workspace Chat shows Decision Lab results as equally ranked bets with readable reasons, linked ideas and accounts, and one confidence basis. Expand a bet to review its evidence and choose Review gaps, Draft opportunity or Drop. Close calls appear under Set aside. Follow-ups enter the composer for you to review and send; the interface does not commit a decision for you.
Inspect source and Context details
Discovery reviews can show a source's native rating and date, variant and provenance. A quote links to a contiguous original passage only when that source text is available; this does not verify generated wording. Workspace Context reads return full scoped bodies and mark any omissions. The assistant must fully read an omitted unit before it can call the context complete. Review cited sources because fuller retrieval does not prove an answer is correct.
Set up workspaces and recover connections
First-run guidance supports typed or spoken setup. Workspace setup is available through the API and CLI. Connected imports can recover a broken connection in place while keeping its known context. Viewer consent and saved-import key grants follow explicit scopes, and CLI reads respect server retry guidance.
Keep Studies independent from Ideas Portal
Teams can run Studies independently of an enabled Ideas Portal, with separate branding and friendly participant addresses. Existing links and workspace sign-in checks continue to work. This release did not test a customer identity provider.
More in this release
Home and planning
- The Home map remains optional; the native Home page stays the default.
- Decision Lab keeps within its time budget and makes its ranked actions and comparison reasons easier to review.
- Huddle and evidence models are updated. Members still review any proposed plan change before it is saved.
Assistant and documents
- Assistant turns record their outcomes, and answer typography stays with its answer.
- Save a tailored document template without supplying a name or description.
Research and public labs
- Compare text models in Public Labs while Responses stays off by default.
- Research models and Study presentation have been updated. Review generated findings before relying on them.
Reliability
- Onboarding and drafting models are updated, and stalled provider attempts can retry within the request time limit.
- Connection and database recovery has improved. Deterministic job failures stop instead of repeating.