You can also browse, search and filter all release notes directly in the application → Open release notes in TAI.
Run any report on a cadence and get it emailed
Major New Feature.
Set up a report once and let TAI deliver it — daily, weekly, monthly, or on any custom cadence, with the same options you know from your calendar (repeat every N weeks, specific weekdays, end dates).
Schedule any report agent from the Reports → Scheduled page, or straight from the run form via Schedule this report.
Email delivery with the PDF attached — pick the recipients and the email language; every delivery is tracked per recipient.
Stay in control: pause and resume schedules, trigger a run on demand, and review the full run and delivery history.
Company admins can restrict who may receive scheduled reports under Settings → Report delivery.
Name a report, set its cadence, add recipients — and it's live
See every schedule's cadence, next run, and status at a glance
Custom recurrence — repeat every N days, weeks, months, or years
Every run and every delivery, per recipient, in one place
Planner & Customer Care Copilot activity will appear in your Analytics
New Feature.
The Planner Copilot and Customer Care Copilot run outside TAI today. Once their sessions are imported into TAI, they'll show up in your Analytics alongside your other agent activity — so your Copilot usage lives in one place. Imports run in the background; there's nothing for you to switch on. Because the sessions come from an external system, they appear a couple of days after they happen rather than in real time.
Agents no longer leak their internal reasoning into the answer
Fix.
When a tool-using agent kept thinking after running a tool, that follow-up reasoning could spill into the visible answer instead of staying inside the Show thinking panel — showing raw deliberation, column names, and half-formed thoughts. Agents running at the default reasoning effort now keep their reasoning in the thinking panel, so the answer you see is just the answer. If you've set an agent's reasoning effort to High, this can still happen — we're tracking the underlying cause. Agents that don't specify a reasoning effort now use Medium; set an agent's effort explicitly if you want a different level.