In Beta
This feature is in beta. Contact your CSM to learn more.
AI Scheduler is a scheduling capability that produces a draft schedule based on the location’s recent schedule history, employee availability, and labor rules. A manager triggers generation, reviews the suggested shifts, and approves, edits, or deletes shifts before publishing. Shifts are approved or deleted individually or in bulk. A suggested shift that violates a labor rule shows the violated rule name, and a flagged shift that adds overtime also shows the added hours.
Security
Users must have the following permission to generate and review AI shifts:
Labor → Scheduling → Schedule → Add/Edit/Delete Shifts
These permissions can be added to custom user roles or individual users. The Permission Access report can be used to determine which user roles or users already have these permissions assigned. For more information, see User Setup and Security.
What AI Scheduler Does
AI Scheduler generates shifts for a full week or specific days in a single action. Managers review each suggestion and keep or discard it before publishing.
The manager stays in control at every step. No AI-suggested shift is saved to the schedule automatically. Every suggestion must be approved before it becomes a real shift, either individually, in bulk, or as the first step of publishing the schedule. When a manager approves suggestions in bulk, R365 validates the entire batch before committing any shift to the schedule.
Where Suggested Shifts Come From
AI Scheduler builds a draft from the location's own history, using three sources in order of priority. It uses the first source with enough data available.
1. Schedule history — The location's last 4 weeks of published schedules. For days in those weeks that were never published, draft shifts are used instead. For each job and weekday, AI Scheduler uses the typical number of shifts, along with the most common start times and shift lengths. Schedule history is the main source, used whenever the location has at least 7 days of schedule history.
2. Punch history — Used when there is not enough schedule history. AI Scheduler looks at the location's last 6 weeks of punches and builds shifts from typical clock-in and clock-out times for each job and weekday. At least 7 days of punch history are required.
3. Labor forecast — Used only when the location has neither schedule history nor punch history, such as a newly opened location. Shifts are built from the published labor forecast for the week being scheduled, using the location's labor matrix configuration. Jobs configured with fixed weekly hours have those hours spread across the week. A published labor forecast is required for the dates being scheduled.
Where the labor forecast is used, how demand is distributed depends on the matrix configuration. A custom labor matrix defines demand by job, so AI Scheduler suggests shifts for each job directly. A general labor matrix defines demand for the location as a whole, and AI Scheduler distributes it across jobs based on the location's staffing history, which is less precise.
What AI Scheduler Does Not Change
Existing shifts are left alone. If the week already has shifts, added manually, from a template, or copied from a prior week, AI Scheduler keeps them and only suggests shifts to fill what is missing. Days that have already passed are skipped.
How Employees Are Matched to Shifts
AI Scheduler reviews each employee's last 8 weeks of published shifts at the location to learn their usual weekly hours and the jobs, days, and times they typically work. Suggestions aim to keep employees close to their usual hours and familiar shifts, while respecting availability, approved time off, job qualifications, and the location's labor rules.
Pending Shifts
Pending shifts are AI-generated suggestions. They appear with a star badge and cannot be moved or resized. A pending shift is not a real schedule entry until a manager approves it.
Pending shifts persist across page reloads. If a manager generates suggestions, closes the browser, and returns to the schedule, the suggestions are still there.
The AI Schedule Panel displays the number of suggested shifts to review, how many could not be filled, and the bulk action buttons.
A pending shift that violates a labor rule shows a warning triangle beside its duration. Pointing to the warning triangle displays the name of the violated labor rule. A shift that adds overtime also shows the added hours in the tooltip, for example 4h daily overtime. A pending shift with no violation shows no warning triangle.
Unfilled Shifts
Some pending shifts cannot be matched to an employee. These shifts appear as a distinct warning-toned card with a plain-language explanation that no employee could be staffed for the shift. Unfilled shifts remain part of the pending count and are not reported as generation failures. Approving an unfilled shift, individually or in bulk, creates an open, unassigned shift on the schedule. The open shift still needs an employee assigned before the week is fully staffed.
Reviewing and Acting on AI-Suggested Shifts
Once generation completes, the Schedule Calendar shows the pending shift cards and the panel transitions to a review view. Managers can approve, edit, and delete suggested shifts individually. Managers can also approve, delete, or regenerate all pending suggestions in bulk. Bulk approval applies to every pending suggestion at once; there is no option to approve a subset by day or job. Learn more about reviewing AI-suggested shifts.
When a manager approves suggestions in bulk, R365 checks the whole batch first and shows a summary before saving anything. The summary states how many suggestions have no warnings, then lists each flagged suggestion with the labor rule it violates. The manager can approve everything listed, skip the flagged ones and save only the rest, or cancel. Skipped suggestions stay as pending shifts and can be approved individually.
AI Scheduler and the Publish Workflow
Approved AI shifts are treated as standard unpublished shifts and go through the normal publish workflow. They are visible to employees only after a manager publishes the schedule.
Publishing a schedule that still has pending suggestions approves those suggestions first, then publishes them. If any of them could not be filled, the publish dialog offers the option to include the resulting unassigned shifts, along with a count of them. On the Employee tab the count covers all unfilled suggestions; on the Job and Department tabs it covers only those for the selected job.
Learn more about publishing a schedule.
FAQ
I approved an unfilled shift and now there's an open, unassigned shift on the schedule — what do I do with it?
An open, unassigned shift still needs an employee assigned before that shift is fully staffed. Assign one the same way as any other unassigned shift on the Schedule Calendar — by dragging an eligible employee onto the shift, pasting a shift onto their line, or opening the shift's details and selecting an employee from the dropdown. Learn more about assigning employees to shifts.
Are "AI-generated shifts" and "AI-suggested shifts" the same thing?
Yes. Both terms refer to the same pending shift suggestions produced by AI shift generation.
Why does a pending shift show a warning triangle?
A warning triangle means the suggested shift violates a labor rule. Pointing to the triangle shows the name of the violated rule, and an overtime violation also shows the added hours. Learn more about labor rules.
Where do suggested shifts come from?
AI Scheduler uses the location's last 4 weeks of published schedules as its main source. If there is not enough schedule history, it uses the last 6 weeks of punch history instead. If the location has neither, it builds shifts from the published labor forecast using the labor matrix.