A woman monitors an equipment scheduling dashboard in an office overlooking an industrial facility.
Microsoft plans to add a planned-maintenance demand forecast to Dynamics 365 Field Service. It will show planners which maintenance events are expected across their assets within a future window they choose, before any work orders exist. Microsoft's AI at Work roadmap (ID 571668) lists a public preview for October 2026 and general availability for February 2027. The feature targets a common problem in field service. Planned maintenance is usually predictable, but until a work order shows up, dispatchers and parts managers have little to plan against. The forecast is meant to move that point earlier, so teams can see where labor, skills, parts and scheduling demand will pile up while there is still time to act.

Dynamics 365 Field Service Puts a Forecast Ahead of the Work Order​

The roadmap entry, titled "Dynamics 365 Field Service: Forecast planned maintenance demand," is marked In development. It lists Android, Desktop, iOS and Web as platforms and the Worldwide (Standard Multi-Tenant) cloud as the only cloud instance. Nothing in the entry covers government or sovereign clouds, so organizations in those environments shouldn't assume the dates apply to them.

Microsoft states the business case plainly. According to the roadmap, maintenance demand often becomes visible to service operations only after work orders are created, which leaves little time to line up capacity and materials. The forecast is supposed to give earlier visibility into expected work across the whole asset population, so planners can spot periods of concentrated demand before they become resource conflicts, parts shortages or delayed maintenance.

The key design point is timing. Microsoft says the forecast "provides visibility before work orders are generated." It is a planning view that sits upstream of the existing work-order, booking and dispatch pipeline. The roadmap doesn't describe it as a replacement for that pipeline or as a way to create work automatically.

Four Maintenance Signals Feed the Field Service Forecast​

The roadmap says the forecast will combine expected maintenance events from four sources:

  • Calendar schedules, meaning maintenance that falls due on fixed dates or intervals.
  • Actual meter usage, where logged readings show that an asset has reached a service threshold.
  • Sequential maintenance plans, where one activity follows another in a defined order.
  • Projected usage based on meter history, which estimates when an asset will reach a usage threshold it hasn't hit yet.

The fourth source is the one that goes beyond simple bookkeeping. Calendar-based maintenance is easy to project, because the dates are known. Usage-based maintenance is harder: an engine due for service at a set number of runtime hours has no fixed date until someone estimates how fast it accumulates hours. By projecting usage from meter history, Microsoft is saying the forecast will turn usage-triggered maintenance into expected dates that can sit on the same timeline as calendar work. The roadmap doesn't name the projection method, and Microsoft hasn't published any accuracy figures.

Planners will be able to see when maintenance is expected and which assets, plans and activities contribute to the forecast. Microsoft also says expected activity duration and other planning details from the maintenance definition "can be used to represent the likely operational demand." In practice, that turns a list of due events into an estimate of work hours, which is the number a capacity planner needs. The forecast can be refreshed as schedules, meter readings and usage predictions change.

Several things are not specified. The roadmap gives no fixed range for the planning horizon, no forecasting algorithm, no accuracy guarantee, no screenshots or UI description, and no setup procedure. It also doesn't say the forecast will reserve parts, assign technicians, create bookings or convert forecast events into work orders. Anyone evaluating the preview should treat those as open questions, not features.

Agreement Booking Dates Show the Gap the Forecast Targets​

Field Service already handles recurring maintenance, so it helps to know what exists today. Microsoft's product overview describes a preventive, agreement-based model in which scheduled maintenance happens on a recurring basis through agreements, and the system automatically generates work orders based on predefined intervals. It sits alongside a reactive break-fix model and a predictive one, in which connected devices detect anomalies and automatically create alerts or work orders through Connected Field Service.

The agreement workflow, as documented on Microsoft Learn, shows the current visibility gap. When an agreement is active and an Agreement Booking Setup (ABS) has a recurrence pattern, Field Service creates Agreement Booking Dates (ABDs) for each scheduled service date. As each date approaches, the system creates a work order based on the ABS's Generate Work Orders Days in Advance setting. If that field is blank, the environment-wide default in Field Service Settings applies. If that is also blank, work orders are generated seven days ahead.

Until then, a booking date keeps the status Active, meaning the date is upcoming and no work order exists yet. Once the work order is generated, the status changes to Processed. Generated work orders enter the Work Orders view as Open-Unscheduled. From there they're booked manually through the schedule assistant, from the schedule board, or automatically by Resource Scheduling Optimization if it's configured. If Auto Generate Booking is enabled on the ABS, the work order is booked to the preferred resource at the preferred start time.

So the future work exists in the system as booking dates, but it only becomes schedulable demand a few days before it's due, spread across individual agreements. Usage-based maintenance faces a similar problem. Rand Group, a Microsoft partner, wrote in May 2026 that when maintenance occurs earlier or later than planned, the native system doesn't automatically reschedule future work orders, and says it has built customizations to adjust future booking dates and add asset meters. That is one partner's account of the platform. It suggests some organizations already build custom tools to get the forward view this roadmap item describes.

The roadmap doesn't say whether agreement-generated work will appear in the forecast. Its four listed inputs are maintenance-plan and meter concepts, not agreements. Organizations that run all preventive work through agreements should confirm during the preview whether that work shows up.

The October 2026 Preview Arrives Through Microsoft's New AI at Work Roadmap​

This feature is one of the first Field Service items announced through a new channel. Microsoft's release-plan pages now say that release plans will no longer be published from September 2026. New Dynamics 365, Power Platform and Dataverse capabilities go to the AI at Work roadmap instead, and the older release plans stay online for reference. Anyone who used to follow the twice-yearly Field Service release-wave documents will need to watch the roadmap for items like this one.

The feature also fits the direction Microsoft set in the 2026 release wave 1 plan for Field Service. That plan listed investments in asset management, maintenance planning and asset location intended to strengthen asset context and "support more proactive planning." The same plan dates the Scheduling Operations Agent's "Automate optimizations" capability to a September 2026 preview and a March 2027 general availability. The maintenance forecast deals with demand and the agent deals with scheduling. Microsoft hasn't said whether the two will be connected.

Microsoft's own caveats apply to the dates. The AI at Work roadmap says its dates are estimates, that all information is subject to change, and that entries are removed when a feature reaches general availability or is cancelled or postponed. The listing is a firm statement of intent, not a delivery guarantee.

ItemRoadmap status
Roadmap ID571668
StatusIn development
Public previewOctober 2026
General availabilityFebruary 2027
PlatformsAndroid, Desktop, iOS, Web
Cloud instanceWorldwide (Standard Multi-Tenant)

The listed platforms include the mobile operating systems, but the roadmap doesn't say what the experience will look like on each. A capacity-planning view is mainly a back-office tool, so it's unclear what a technician on the Field Service Mobile app would see, if anything.

What Field Service Teams Should Do Before the Preview​

If you're a Field Service administrator or maintenance planner, the useful work now is getting your data ready. The forecast can only project from maintenance definitions, schedules and meter readings that already exist in the system. If those are incomplete, the forecast will be too, and a preview will tell you little.

Start with the inputs Microsoft has named. Check that assets that should be on a maintenance plan actually have one, and that calendar and usage triggers are set up the way your operation really runs. Meter data deserves attention, because the usage projection depends on a reading history. Assets with irregular or missing readings give a projection little to work from. Microsoft says expected activity duration from the maintenance definition feeds the demand estimate, so blank or placeholder durations will distort the labor picture.

Teams that rely mainly on agreements should also know where their current forward view comes from. The Active booking dates under each Agreement Booking Setup, together with the days-in-advance setting, define how much notice dispatchers get today. That gives you a baseline for judging whether the preview adds anything.

  • The planned-maintenance demand forecast is in development, with a public preview listed for October 2026 and general availability for February 2027. Both dates are Microsoft estimates.
  • Only the Worldwide (Standard Multi-Tenant) cloud is listed, so other cloud environments shouldn't plan around these dates.
  • The forecast is designed to combine calendar schedules, actual meter usage, sequential maintenance plans and projected usage from meter history. Maintenance-plan and meter data quality will limit how useful it is.
  • Expected activity duration from maintenance definitions feeds the demand estimate, so fill in realistic durations before testing.
  • Microsoft hasn't said the forecast will create work orders, bookings or parts reservations. Existing agreement generation, Resource Scheduling Optimization and custom date-adjustment logic keep their current roles.
  • Field Service feature announcements now go to the AI at Work roadmap instead of release-wave plans, so track roadmap ID 571668 there for changes to status and dates.

For organizations that plan maintenance weeks or months ahead, this is one of the more practical items on the Field Service roadmap. It turns maintenance data that already sits in Dynamics 365 into a forward view of labor, skills and parts demand, instead of work that appears about a week before it's due. The October preview is the first chance to see how far ahead the planning horizon reaches, how well the usage projections hold up, and whether agreement-based work is included. Until February 2027, the forecast remains a roadmap commitment, and the best preparation is clean maintenance plans and reliable meter readings.