Save assistant work once
Instructions, sources, model policy, and checkpoints persist. Relam runs in the background and returns a result you can inspect—not a thread you have to find again.
Relam features
Schedules & remindersUseful recurring work should not depend on remembering the same prompt.
Relam Schedule saves assistant tasks—instructions, sources, timing, and checkpoints—so briefings, inbox attention, and monitors run whether you are in the room or not. Confirm once. Inspect run history when they finish.
The problem everyone has
Useful recurring work should not depend on remembering the same prompt.
Teams already know what should run again—inbox attention before standup, a weekly status from the same sources, a monitor when a client thread moves. The work fails when it lives only in someone's head or at the bottom of a chat.
- The same Monday brief gets rebuilt from Slack, mail, and docs.
- Inbox triage waits until you are back at your desk.
- Useful prompts disappear when the chat thread scrolls away.
How Relam Schedule solves it
Give useful work a rhythm—with sources, timing, and history built in.
Put it on a rhythm or an event
Weekday inbox attention, weekly project status, or a run when matching mail arrives. Timing respects your timezone; overlap policies prevent stacked duplicates.
Return to run history
Successes, failures, deliveries, and no-change results live in Schedule. Your team delegates with evidence—not “did anyone check this yet?”
Reminders when that is enough
A timed nudge with your saved note when you do not need a model run or source reads.
The Schedule workspace—list, editor, sources, and run history.
Scroll inside the editor panel to review prompt, frequency, sources, and notifications—the same fields you confirm before a task saves. Tab through assistant, recurring, and event-triggered examples without the page jumping.
What changes when the task runs whether you are in the room or not.
Leaders using scheduled assistant tasks describe what changed when recurring work left chat history and moved into Schedule—with sources, cadence, checkpoints, and honest run history.
Jamel AbedMicrosoft MVP
“Relam Schedule turned the prompts I kept rewriting every week into saved assistant tasks—with checkpointed sources and run history I can actually inspect.”
Sophia R.CVP, Skytells
“We started adopting Schedule for cross-region handoffs—repeatable assistant work with scoped sources so the next shift inherits a result, not a blank thread.”
Enterprise-ready means honest requirements—not hidden surprises.
Schedule is governed like the rest of Relam: plan limits, scoped connector reads, preview before save, and run history you can audit. Here is what the product actually checks before a task runs.
Compare plan limits- Schedule on your plan
Your plan must include Schedule with active task capacity and monthly compute units. Limits appear before you save.
Active schedules · cadence · units per run - Assistant task prerequisites
Background-capable model access, attached instructions, and optional library, project, or connector sources you authorize.
Preview timing and task type before the first run - Connected mail or calendar
Inbox-attention and meeting-prep tasks need Gmail, Outlook, or calendar plugins connected with the scopes you approved.
Reads are bounded and checkpointed—never whole-account access - Event triggers
Polling mail and calendar for new or updated records requires an eligible plan with event triggers enabled on the deployment.
Activation does not replay historical backlog - Delivery you configure
Notification channels and digest preferences are set per task. Sensitive previews stay off by default.
In-app inbox · email · push where enabled - Workflow tasks (optional)
Skytells workflow runs need your own connection, approved workflow ID, and a plan that allows workflow schedules.
Started does not mean completed—history stays honest
Your inbox and calendar are inputs—not separate apps to check first.
Assistant tasks can include authorized mail and calendar connectors alongside library and project context. Relam reads within the scopes you approved, checkpoints what it saw, and reports when a required source failed—never pretending nothing changed.
GmailInbox-attention tasks read authorized mail since the last successful run
Outlook EmailSame for Microsoft mail—bounded queries, checkpointed reads, cited results
Google CalendarMeeting-prep tasks see upcoming events in the week you already live in
Outlook CalendarCalendar-aware tasks stay aligned with Outlook meetings—not a second schedule
Connect accounts in Plugins. Assistant tasks reference them as connector sources—bounded reads with checkpoints, not blanket access to your whole mailbox or calendar.
How Navix automates operations
and delegates with Schedule.
Navix runs international logistics across regions and time zones. Schedule turns repeatable assistant work—inbox attention, lane status, client monitors—into saved tasks with timing, sources, and run history the next shift can inspect.
We schedule inbox attention, lane status, and client-thread monitors so operations keep moving across regions—without asking everyone to rebuild the same brief every Monday.
Inbox attention before handoff
Regional leads get a weekday assistant run on authorized mail—attention list in Schedule before the next timezone takes the lane.
Weekly lane status without rebuilding
Project instructions and pinned library sources feed a Monday task. The team reviews only when checkpoints show something moved.
Client threads on events, not guesswork
Event triggers on matching mail queue a summary when a client thread changes—delegation with evidence, not a Slack ping to “check email.”
Repeatable work stops living in your chat history.
Chat is for thinking together. Schedule is for work that should happen again—briefings, monitors, project status, document change checks—with model runs, checkpoints, and notifications when something actually needs you.
Assistant tasks that run without you
Save instructions, pick a model, and let Relam execute in the background—reading only the sources you attached.
Library and project context included
Pin documents, attach a project brief, or continue a chat. Scheduled runs reuse the same context boundaries as the rest of Relam.
Connected mail and calendar reads
Authorized Gmail, Outlook, and calendar accounts feed inbox-attention and meeting-prep tasks—with checkpointed, bounded reads.
Once, recurring, or on events
One-time runs, RRULE/cron cadences in your timezone, or event triggers when new mail or calendar records appear on eligible plans.
Runs you can audit
History, failures, deliveries, and results live in Schedule—not buried in a chat you already moved on from.
Reminder
When you only need a nudge, not a model run. Reminders surface your saved note at the chosen time—no source reads, no background assistant execution.
Assistant tasks, event triggers, workflow runs, and compute limits vary by plan and deployment settings. Compare current plans.
Questions about Schedule, answered plainly.
Assistant tasks, reminders, event triggers, and plan limits are governed the same way in the product and on this page—no hidden surprises after you save.
What is the difference between an assistant task and a reminder?
Assistant tasks are background model runs with your instructions, optional sources, checkpoints, and run history. Reminders are timed nudges that surface your saved note—no source reads and no background assistant execution.
Does Schedule send email or post on my behalf?
No. Scheduled assistant tasks read authorized sources and return results inside Relam. They do not send mail, post to Slack, or write to connected tools without the same review flows you use in Work.
What sources can an assistant task read?
Library documents, a project brief, public URLs, an optional chat continuation, and—when connected—authorized Gmail, Outlook, or calendar accounts within the scopes you approved.
What are event triggers?
Event triggers poll connected mail or calendar for new or updated records. When a match appears, Relam queues an assistant run with bounded, checkpointed reads. They require an eligible plan and do not replay historical backlog.
Can my team see what ran?
Run history, failures, deliveries, and checkpoints live in Schedule per account. Project context can be shared through Relam projects—the task definition and its audit trail stay where you saved them.
What plan limits apply?
Active schedules, cadence, compute units per run, event triggers, and workflow tasks vary by plan and deployment settings. Limits are surfaced before you save a task.
Where do I compare plan limits?
Active schedules, compute units, event triggers, and workflow tasks vary by plan. Compare current plans or open Schedule in your workspace to see limits before you save.
Governed AI. Personal boundaries. Multi-model freedom.
Privacy at Relam