Dynamics 365 Field Service Turns Meter Readings Into a Planning Date
Usage-based maintenance ties service to the work an asset has actually done, such as run hours, cycles or distance, instead of a fixed calendar interval. The weakness is timing. Under a meter rule, the threshold is reached whenever it is reached. Unless someone works out the asset's pace by hand, the organization finds out at the trigger point or just before it.
Roadmap item 571667 addresses this with a simple method. Field Service would calculate a projected usage rate for each asset from the mean of its meter-usage logs over the trailing 30 days. That one rate represents the asset's recent pace and drives the projected due dates for every usage-based maintenance activity on that asset.
For each usage-based requirement, the system divides the usage remaining before the next threshold by that rate and converts the result into a calendar date. Here is an illustration, not a Microsoft example. An asset 400 run-hours short of its next service that has averaged 8 hours a day over the past month would be projected about 50 days out. If the asset runs harder or sits idle, the projection moves as new readings come in.
The roadmap description is careful about one boundary. The configured meter threshold is still the maintenance requirement. The projected date is an advance planning signal that feeds maintenance forecasts and lead-window work-order generation. Nothing in the description says the forecast replaces the meter rule or overrides it.
Why a 30-Day Mean Is a Deliberate, Explainable Choice
Microsoft is not presenting this as a machine-learning failure model. The method is a trailing average, and the roadmap text says the approach is meant to stay "explainable and reviewable by planners." A dispatcher or reliability engineer can check the number by looking at the asset's last month of readings. That is a real advantage when a projected date will be used to commit technicians or plan customer downtime.
It also sets limits on what the feature can do. A 30-day mean describes recent behavior. It cannot anticipate a seasonal ramp, a shift change or a customer who is about to double an asset's workload. When operating patterns change, the forecast lags until the new pattern has worked its way through the 30-day window. That is the cost of choosing a simple, auditable number over a more elaborate model.
Microsoft has added two safeguards. The first is outlier protection. Field Service is supposed to stop isolated meter spikes, wrong readings and other anomalous data from pulling the rate out of line. Anyone who has seen a technician type an extra zero into a meter field knows why this matters. The second is ongoing recalculation. The rate is recomputed as more meter history arrives, so later predictions keep tracking recent usage.
The roadmap text does not say which outlier algorithm is used, how much history an asset needs before its projection counts as reliable, how often recalculation runs, or whether uncertainty is shown as anything more than a single date. Teams planning to trust the projection should evaluate those points in the preview.
Stored Rate and Timestamp Give Planners an Audit Trail
The item makes one design choice worth noting. Field Service would store both the calculated usage rate and the date and time it was last calculated. Microsoft's stated aim is to let planners judge how fresh the estimate is and whether there is enough history behind it to trust the projection.
This works as a practical freshness check, not a formal confidence score. A projection built from a rate calculated yesterday on a steady stream of readings deserves more weight than one built from a stale rate on an asset that logged two readings last month. With both values visible, planners can see the difference without exporting data.
The stored values also help administrators. Because the rate and timestamp are persisted, it is reasonable to expect them to be available for views, reports and automation like other Field Service data. That is an inference: the roadmap does not describe any Power BI, Power Automate or API exposure. Still, stored values are generally easier to build on than a figure calculated only for display.
Where Roadmap Item 571667 Fits Microsoft's Broader Maintenance Rework
The forecasting item is part of a wider change to preventive maintenance in Field Service. Microsoft's 2026 release wave 1 plan includes a related feature, "Predict asset service dates by usage or calendar," which Microsoft describes as a way to plan preventive service with usage, calendar, and milestone triggers, plus active windows, sequencing, and manual overrides per asset. The two are separate entries with their own schedules. The research notes for this story say the release-plan feature defines trigger types and due-date recalculation, while item 571667 specifies how a usage rate is calculated and turned into a date. Readers should not assume the two share a release date.
Before this work, usage-based maintenance in Field Service often depended on partner extensions or custom builds. Microsoft partner HSO wrote in October 2025 that usage-based service until recently it wasn't something you could easily set up in Field Service, and that its own add-on lets service managers track asset properties—like airflow, mileage, or run-time hours—and trigger a work order when those thresholds are reached. Rand Group, another partner, describes customizing the product by adding Asset Meters to Dynamics 365. Its example has a construction company alerted as an engine approaches 900 hours of runtime so the work can be scheduled before the 1,000-hour service mark.
Those partner approaches show what planners have asked for: warning before the threshold arrives. They mostly do it with fixed lead offsets on the meter itself, such as alerting 100 hours early. Item 571667 does it in calendar time, derived from how fast the individual asset is actually running. For a planner booking technicians by the week, a date is easier to act on than a meter margin.
The item's timing also reflects a change in how Microsoft publishes plans. Microsoft's Field Service release-plan pages now say release plans will stop being published from September 2026. New Dynamics 365, Power Platform and Dataverse capabilities are moving to the AI at Work roadmap, which is why this Field Service feature shows up next to Microsoft 365 Copilot announcements. Admins who tracked Field Service changes through Microsoft Learn release plans should add the AI at Work roadmap to their watch list.
How the Forecast Connects to Work Orders and Scheduling
A projected date only helps if it reaches the scheduling process. In Field Service that process runs through work orders. Microsoft's documentation says work orders are used to coordinate and schedule resources and activities, including installations, repairs and preventive maintenance. After creation, a work order is scheduled manually, with the schedule assistant, or with the Resource Scheduling Optimization add-in. When the work is done, a supervisor reviews and approves it.
The roadmap says the projected date feeds "lead-window work-order generation." In plain terms, work orders could be created some period before the forecast due date instead of at the moment the meter crosses the threshold. The description does not say how long that window is, whether it can be configured, or whether generated work orders are booked automatically. The roadmap supports the claim that the forecast drives work-order generation within a lead window. It does not support automatic dispatch, and planners should not assume it.
The inputs deserve attention too. The forecast is only as good as the meter log behind it. Organizations that record meter readings by hand during service visits will produce sparse, uneven histories, and a 30-day mean over two data points is weak evidence. Organizations whose connected equipment writes usage directly into the asset's meter log will have much denser input. The research notes point out that Microsoft's documentation for the related maintenance capability describes connected devices keeping projected dates current this way. The roadmap entry for 571667 does not list the integrations or setup it requires.
What this means for you
If you run preventive maintenance in Dynamics 365 Field Service, spend the time before the October 2026 preview checking your meter data, not changing configuration. The feature does not exist in production tenants yet, and its dates are estimates. How useful it turns out to be depends on whether your assets already have steady, accurate usage logs.
Organizations still running usage-based maintenance on a partner extension or custom meter entities face a more specific choice. They will need to decide whether the native feature can replace that customization, and they can only judge that properly during the preview, once the undocumented details are visible. Teams running only calendar-based maintenance can wait, because this item applies to usage-based requirements.
- The forecast is planned, not shipped: roadmap ID 571667 lists public preview in October 2026 and general availability in February 2027, and only for the Worldwide (Standard Multi-Tenant) cloud.
- The configured meter threshold remains the maintenance requirement, and the projected date is a planning signal derived from a trailing 30-day mean of the asset's meter-usage logs.
- Audit your meter data now: assets with sparse, manually entered or error-prone readings will produce weaker projections, even with outlier protection.
- Use the stored usage rate and last-calculated timestamp to judge whether a projection is fresh enough to plan labor, parts or customer downtime around.
- In preview, check the details the roadmap leaves open: minimum history, recalculation frequency, the outlier method, and how lead-window work-order generation is configured.
- Watch the AI at Work roadmap for changes, because Microsoft is moving Dynamics 365 plans there from the retired Microsoft Learn release plans.
Microsoft is giving meter-based maintenance something it has lacked in the standard product: a date planners can act on, calculated in a way they can check for themselves. Choosing a trailing average and displaying the stored rate and timestamp means planners can see why a date was projected. The limits come from the same simplicity, since the forecast only knows the last 30 days of readings. The October 2026 preview is the point at which Field Service administrators can find out whether their meter history is dense and clean enough to rely on, ahead of the February 2027 general availability target.