Predictive work duration
Field service dispatchers often assign a fixed duration for a job type, despite various factors that might impact the time that is required to complete the work. For example, a work order to perform system maintenance might be estimated to take an hour. A less experienced technician might take more than an hour, whereas a more experienced technician might be able to complete the ticket in half that time. Using fixed durations could cause delays that can cascade across the whole company, potentially lowering use and reducing the overall effectiveness of scheduling.
By analyzing historical data and developing a better understanding of past outcomes, organizations can make more informed predictions about the amount of time that an item will take to complete. The Predictive work duration feature in Dynamics 365 Field Service can help predict scheduling outcomes, which can improve overall efficiency.
By using AI models, the Predictive work duration feature learns from historical booking information and completion times. Based on factors such as past resource performance, the customer, and the type of work, this feature can compute a more realistic duration. Dispatchers can use this computed duration to decide how an item should be scheduled.
Factors that influence predictions include:
Resource performance
Incident type
Customer (service account)
Work order type
Territorial differences
Time-related changes
Get started
Before you can start applying predictive work durations, you'll need to enable the feature in your environment. In the Dynamics 365 Field Service application, go to the Settings area and then select Settings under the Analytics and insights section. Select Status on the Predictive work duration (preview) page.
After you have enabled the feature, you should decide if you want to enter a value into the Minimum number of resources in a territory field. This field helps you set restrictions on analyses and reports. Territories that have the same or fewer resources than the value that you have defined in this field will be omitted. For example, if you set this field to 10, then territories that have 10 or fewer resources will be excluded from territory analyses and reports. This limitation can help protect technician privacy.
When you have enabled the Predictive work duration feature, you'll have access to three reports:
Upcoming work duration - Shows predicted durations for upcoming bookings and compares the predictions to the allocated amount.
Territory duration - Provides prediction breakdowns by territory.
Resource duration - Provides prediction and allocation details by resource.
The following sections examine each report in more detail.
Upcoming work duration
The Predictive duration report is available from the Analytics and Insights group in the Resources area of the Field Service application. Within the Predictive duration report, the Upcoming work duration tab includes two charts and provides predictions on unscheduled requirements and scheduled bookings.
The upper portion of the report provides ways to filter or slice the data:
Date - Select the future date (next seven days).
Territory - Select from the list of resource territories.
Incident type - Select from the list of incident types.
The following analytics are beneath the slicers:
Under allocated bookings - Number of bookings that are allocated under predicted duration.
Under allocated requirements - Number of requirements that are allocated under predicted duration.
Each chart includes the following metrics:
Allocated/Requirement duration - Specifies how time is currently allocated to a scheduled booking or defined on the requirement record.
Predicted duration - Defines how much time the system predicts that the job will take based on different factors.
Difference - The total difference between the Allocated/Requirement duration and the predicted duration.
Confidence - Likelihood that the actual duration will be close to the predicted duration. A higher score (closer to 100 percent) signifies a better chance of the prediction being correct, while a lower score (close to zero percent) means that the prediction might be incorrect.
Prediction factors - Insights into factors that were considered when you were predicting the duration and how those factors affect the prediction.
Territory duration
The Territory duration tab provides a report that separates details by territories.
As with the Upcoming work duration report, the Territory duration report includes multiple ways that the information can be filtered based on your needs:
Duration - Lists predefined values to filter by duration, such as:
Last three months
Last six months
Last one year
Last two years
Incident type - Filters by incident types.
Territory - Filters by service territories.
Customer - Filters the territory report by customer.
Beneath the filters are analytics that will highlight important measurable elements that are related to your organization:
Number of resources - Shows the number of resources that are involved in bookings.
Number of bookings - Shows the number of completed bookings.
Overestimated bookings - Shows the number of bookings that are overestimated compared with actual duration.
Underestimated bookings - Shows the number of bookings that are underestimated compared with actual duration.
Overall proficiency score - Overall AI proficiency across all territories.
The Territory duration breakdown report includes the following charts to help simplify the process of breaking down and visualizing information:
Booking volume - Shows the number of bookings over the chosen filters (duration, territory, and so on).
Booking duration - Compares overestimated and underestimated bookings over the chosen filters (duration, territory, and so on).
Customer metrics - Shows bookings, difference, and the overestimates and underestimates by customer.
Incident type metrics - Shows jobs completed, estimated duration, average duration, difference, proficiency score, confidence, and proficiency factors by incident type.
Territory - Map that visualizes metrics by defined territories.
Territory metrics - Shows jobs completed, proficiency scores, confidence, and proficiency factors for each territory.
Resource duration
You can find the Resource duration report by going to Resources > Analytics and Insights > Resource duration.
The Resource duration report includes several analytics to help show the factors that impact your organization. The upper portion of the report shows the following analytics:
Number of resources - Shows the number of resources that are involved in bookings.
Number of bookings - Shows the number of completed bookings.
Overestimated bookings - Shows the number of bookings that are overestimated compared with actual duration.
Underestimated bookings - Shows the number of bookings that are underestimated compared with actual duration.
In addition to the preceding analytics, the Resource duration report includes the following charts that dispatchers can use to better understand your organization's bookings over a defined period:
Booking volume - Shows the number of bookings over the chosen period.
Booking duration - Compares overestimated and underestimated bookings over the period.
Resource metrics - Shows the number of jobs that are completed by a resource.
Incident type metrics - Shows jobs completed, estimated duration, average duration, difference, proficiency score, and proficiency factors by incident type.
Each metrics chat contains more detailed metrics that define the breakdown of the score:
Proficiency score - A score that indicates the likelihood of completing a booking within the estimated duration. A higher score indicates a high likelihood of completing a booking within the estimated duration and a lower score indicates the opposite.
Confidence - Likelihood that the proficiency score that is calculated from the samples is close to the true proficiency score. A higher score indicates the proficiency score being closer to the truth, and a lower score indicates the proficiency score being further from the truth.
Proficiency factors - Summary of factors such as incident types, service accounts, day of the week, and the skills and rating match that contribute to the efficiency score.
For more information, see Predictive work durations and resource proficiency suggestions.