Applying Artificial Intelligence 

Objective

After completing this lesson, you will be able to apply AI features in SAP Field Service Management to enhance scheduling, filtering, and summarizing operations.

Introduction to Artificial Intelligence

Unit 10: Key Topics

In the next segment of this course, you will uncover the potential of integrations between artificial intelligence (AI) and SAP Field Service Management (FSM). Here are some focuses and objectives as you progress:

  1. Overview of AI in FSM: Get introduced to the concept of using AI in FSM and how it enhances the field service experience.

  2. Definitions of AI Functionalities: Understand the difference between Base and Premium AI functionality and learn about the licensing requirements for using SAP AI Units.

  3. In-depth look at AI Features: Dig into each AI feature available in FSM, which includes Assignment Duration Prediction (a base feature) and Intelligent Filtering, Equipment Summary, and Activity Summary (premium features). Each feature’s purpose, use, and requirements will be discussed in detail.

  4. Understand Anonymization of Personal Data: Learn how FSM AI filters out personal identifiers when communicating with Large Language Models (LLM) to maintain data privacy.

  5. Learn how to Enable and Use AI Features: Detailed steps on how to enable both base and premium AI features such as Intelligent Filtering, Equipment Summary and Activity Summary will be explained.

By the end of this unit, you should have a solid understanding of how AI can be utilized in FSM to enhance dispatching and technicians' work. You will also understand how to enable these features and ensure your use of AI aligns with data privacy standards.

Overview of AI Features in FSM

SAP Field Service Management is using artificial intelligence (AI) to help dispatchers and technicians to provide the next level field service experience for their end customers.

AI functionalities are available either as Base or Premium AI functionality.

Premium AI functionalities are available only when you have a valid license to use SAP AI Units. When you have the license, you can create a request to activate Premium AI features in SAP Field Service Management. To create the request, raise an incident with the CEC-SRV-FSM-AI component and mention that you have a valid license for SAP AI Units. When the AI features are activated, the GENERATIVE_AI role is assigned to your account and you are informed about the activation. From then, you can start using the premium AI functionalities.

Below, you find an overview of the AI functionalities we currently offer together with their type and the model they are using:

  • Assignment Duration Prediction (Base feature, internal AI model)
  • Intelligent Filtering (Premium feature, generative AI model)
  • Equipment Summary (Premium feature, generative AI model)
  • Activity Summary (Premium feature, generative AI model)

Personal Data Anonymization

When external large language models (LLM) are called, the input is scanned for personal data, for example, person, phone numbers, IP address, and so on. This data is removed from the request to the LLM and replaced with unique identifier tags so that - if included in the response - the actual value can be included again in the result on SAP Field Service Management side.

Assignment Duration Prediction

Assignment duration prediction is a an AI feature for semi and fully automated scheduling in SAP Field Service Management. You can use the feature to let the system predict the assignment duration based on historic data. The purpose of the prediction is to have more a realistic schedule for the technicians that considers skills, customer, equipment, location, and so on.

Displayed Results

If the predicted duration is enabled, you see the predicted value instead of the planned duration. Depending on the AI-based scheduling feature, the results are displayed at different locations. Below, you find an overview.

Image showing the Predicted Duration visible in an Autoscheduling Reports

For automated scheduling, you will see the predicted value in the Auto-Scheduling reports as part of the log messages in the activity details. The predicted value is displayed in the Planned Duration field together with a tick icon in the Planned Duration field.

Image showing the Predicted Duration visible on the Best Matching Technician widget

If you are using the Best Matching Technician feature, the predicted duration is displayed for each technician in the technicians list on the Dispatching Board. From there, you have the option to edit the duration if you are not satisfied with the predicted value. When you assign the activity to a technician, the predicted duration will be used by the system. If you edited the value, the edited duration will be considered.

If you are using the Planning Widget, the predicted duration will be used and the activities will be assigned using the predicted value. You will see the predicted value in the Auto-Scheduling reports as part of the log messages in the activity details.

How the Prediction Works and Used Data

To predict the assignment duration, a machine learning model is used. The model is trained based on the available time efforts and the data of closed activities to determine the real duration of the closed activities. In addition, the model uses the following data of the closed activities during the training to learn which characteristics influence the duration of an assignment:

  • Service call priorities
  • Equipment ID
  • Business partner ID
  • Skills of the responsible person
  • Required skills of the activities

For the model training, we offer a sample business rule. By default, the sample business rule includes the following settings:

  • The training data of one year is used.
  • The model training takes place once a day.

A sample business rule for training the data model is available on the SAP Help Portal.

Limitations

There are a few limitations and tips to consider for the feature:

  • If new data is very different from the past training data, then the estimation could become less useful.
  • The more data you have available, the better for the model training and the output. However, note that for performance reasons the model training uses at maximum of 10,000 closed activities within the provided time range.
  • Currently, we do not consider user-defined fields (UDFs) during the model training.
  • We support training the model once a day. If you train the model more often, the training could fail.

Enabling Assignment Duration Prediction

Before Assignment Duration Prediction is enabled, make sure the following prerequisites are in place:

  • You have 1000 activities with the execution Stage equal to CLOSE.
  • You have time efforts maintained for each closed activity. These should represent the start and end of the activity when it was performed by the technician.
  • It is recommend that you have maintained the following data for the closed activities:
    • Service call priorities
    • Equipment ID
    • Business partner ID
    • Skills of the responsible person
    • Required skills of the activities
  • You have the license "SAP FIELD SERVICE MANAGEMENT, SUPPLEMENTAL SERVICES".

To activate assignment prediction, follow the procedure below:

  1. Configure and train the machine learning model for the assignment duration prediction. For the model training, use the sample business rule provided on the SAP Help Portal.
  2. Activate the Allow assignment duration prediction for semi and fully automated scheduling checkbox in the FSM web UI (shell) under Settings and Configuration → Planning and Dispatching → General settings → Activity Planning → Additional Settings.
  3. The predicted duration will now be applied when using semi- or fully automated scheduling.

Intelligent Filtering

Besides the basic and advanced filters that you can use in the Planning and Dispatching application, there is also the option to use intelligent filtering. Intelligent filtering allows you to filter the data in the planning list using natural language (queries). Based on your query, the data is filtered accordingly. Here is an example:

Image showing Intelligent Filtering on the Planning Board

Please note that when you use intelligent filtering and you have any manual filters applied, the manually applied filters are not considered.

Your input query and the metadata about the object (the kind of attributes it has) are anonymized and then forwarded to a Large Language Model. The model processes the data and creates a filter based on which it filters the data.

The following query tips can help the model to process your query and provide the desired results:

  • Explicitly mention service calls.
  • Use verbs like "Display", "Show", and so on.
  • Explicitly mention what attributes you want to filter by. For example, writing "Display service calls for customer XY" is more effective than "Display service calls for XY".

Limitations

There are a few limitations to consider for this feature:

  • You can only enter your query in English.
  • Your query can have a maximum of 200 characters.
  • Do not provide any sensitive, personal identifiable information in the query.
  • The systems supports a maximum of 10 queries per minute.
  • Intelligent filtering only works for the planning list.
Image showing the re-use previous queries

Previous queries can be re-used because the system automatically saves the queries that you execute. Whenever you start writing a new query, the previously enter queries are displayed in the dropdown. You can choose one of these and it will be executed without consuming additional AI units.

Enabling and using Intelligent Filtering

Before Intelligent Filtering is used, make sure the following prerequisites are in place:

  • You have a license to use SAP AI units.
  • The SAP Premium AI features are activated for your account in SAP Field Service Management.

To use Intelligent Filtering, follow the procedure below:

  1. Open the Planning and Dispatching application.
  2. Go to the Dispatching Board.
  3. Above the planning list, switch to the Intelligent Filtering mode by choosing Natural Language Filter.
  4. Enter your query in the text field.
  5. Choose Submit Query.
  6. The service calls are now filtered according to your query.

Note

Artificial Intelligence generates results based on multiple sources. Outputs may contain errors and inaccuracies. Consider reviewing all generated results and adjust them if necessary.

Equipment Summary

Equipment summary is an AI feature that allows you to generate a summary based on previous performed activities for a piece of equipment. The purpose of the summary is to provide the technician with information about the piece of equipment, for example, used spare parts, the duration of the activity, performed checklists and their results. In this way, the technician can resolve any occurring issues quicker. It is also helpful for the dispatcher to better decide to whom new activities should be assigned. The summary also contains links to relevant information, for example, to the technician details, reserved materials, and so on.

Image showing equipment insights

You can view the summary from the Service Calls screen or the Dispatching Board.

Using Equipment Summary

Before the Equipment Summary is used, make sure the following prerequisites are in place:

  • You have a license to use SAP AI units.
  • The SAP Premium AI features are activated for your account in SAP Field Service Management.

To use the Equipment Summary feature, follow the procedure below:

  1. Open the Planning and Dispatching application.
  2. From the side menu, choose between the following options:
    • Go to the Dispatching Board and choose a service call code from the planning list.
    • Go to the Service Calls screen and choose a service call code from the list.
  3. On the General Information tab, go to the Insights section.
  4. Choose the button to generate the summary.
  5. The equipment summary is now displayed on screen. You have the option to copy the summary for further use or to close it again.

Note

Artificial Intelligence generates results based on multiple sources. Outputs may contain errors and inaccuracies. Consider reviewing all generated results and adjust them if necessary.

Activity Summary

Activity summary is an AI feature that allows you to generate a summary based on the information of the activity. The purpose of the summary is to provide the technician with information about the activity - for example, used spare parts, the duration of the activity, performed checklists and their results. This information can be helpful to perform new activities. The summary will also contain links to the relevant information, for example, the technician details, reserved materials, and so on.

Image showing Activity Insights

To generate the summary, the system collects all activity data, this includes:

  • Checklists
  • Reserved materials
  • Time efforts

The collected data is then forwarded to a Large Language Model that generates the summary of the activity.

Using Activity Summary

Before the Activity Summary is used, make sure the following prerequisites are in place:

  • You have a license to use SAP AI units.
  • The SAP Premium AI features are activated for your account in SAP Field Service Management.

To use the Activity Summary feature, follow the procedure below:

  1. Open the Planning and Dispatching application.
  2. From the side menu, go to the Activities screen and choose an activity from the list.
  3. On the General Information tab, go to the Insights section.
  4. Choose the button to generate the summary.
  5. The activity summary is now displayed on screen. You have the option to copy the summary for further use or to close it again.

Note

Artificial Intelligence generates results based on multiple sources. Outputs may contain errors and inaccuracies. Consider reviewing all generated results and adjust them if necessary.

Challenge Question

Challenge Yourself: Putting Your Knowledge to the Test

In this lesson, you'll have the opportunity to apply the concepts and knowledge you've gained throughout the unit. We've designed an engaging Challenge Question that will put your critical thinking skills to work. Take a moment to reflect on what you've learned, and then use that understanding to craft your own unique solution to the question at hand.

To make the most of this exercise, we encourage you to write down your answer on a separate piece of paper. This will help you organize your thoughts and measure your learning progress. When you've completed your answer, compare it to the expert response provided. This will give you valuable insight into how well you've grasped the material and where you might need to focus your attention for further growth.

Remember, this is an opportunity to apply your understanding in a practical way, so don't hesitate to think creatively and explore different approaches. Your active participation in this lesson will reinforce your learning and prepare you for success in the real world.

Scenario:

You are a Field Service Manager at a large manufacturing company that uses SAP Field Service Management (FSM). Recently, the company has acquired licenses for SAP AI Units and wishes to leverage AI functionality to enhance field service experiences. Your task is to implement the AI features in your existing FSM operations.

Task:

  1. Explain how you would activate and utilize assignment duration prediction for semi and fully automated scheduling.

  2. Describe your strategy for implementing Intelligent Filtering in the planning list.

  3. Detail the steps you would follow to enable and use Equipment and Activity summary features.

  4. Discuss the importance of personal data anonymization in AI features and how SAP FSM ensures data privacy.

Expert Consultant Response

  1. For activating Assignment Duration Prediction, I would first ensure there are at least 1000 closed activities with corresponding time efforts that represent the performed task by the technician. Then, I would activate the "Allow assignment duration prediction" checkbox under Settings and Configuration → Planning and Dispatching → General settings → Activity Planning → Additional Settings in the FSM web UI. Now, predictions from the machine learning model would be used during semi and fully automated scheduling.

  2. To implement Intelligent Filtering, I would switch to the Intelligent Filtering mode in the Dispatching Board's planning list and start entering query in English. This English language query would help the AI model quickly filter data in the planning list based on the entered query.

  3. I would enable the Equipment Summary and Activity Summary features by accessing an individual service call or activity's General Information tab. In the "Insights" section, there's an option to generate the summary. Upon activation, the AI generates a summary of all relevant details to help technicians in handling the task efficiently.

  4. Personal data anonymization is key in AI features as it safeguards sensitive data from potential exposure during external large language model processing. In the SAP FSM, the system scans the input for any personal data when external LLMs are called. This data is removed and replaced with unique identifier tags, ensuring that the actual values are included in the result only on SAP FSM side, thereby ensuring data privacy.

Remember that it's critically important to avoid including any sensitive, personally identifiable information in intelligent filtering queries. This continues to emphasize the importance of privacy and security when leveraging AI features in FSM.

Lesson Recap

This lesson revolved around understanding the implementation and usage of Artificial Intelligence (AI) features within SAP Field Service Management (FSM).

Key topics included:

  1. Overview of AI in FSM: We learned that both base and premium AI functionalities are available in FSM. Premium functionalities require a valid license for SAP AI Units.

  2. AI Features in FSM: We explored several AI features, including their types and models:

    - Assignment Duration Prediction: A base feature using an internal AI model that predicts assignment duration based on historical data.

    - Intelligent Filtering: A premium feature that uses a generative AI model to enhance data filtering capabilities.

    - Equipment Summary and Activity Summary: Premium features that generate summaries based on past activities related to a piece of equipment or an activity, respectively.

  3. Personal Data Anonymization: FSM ensures data privacy by removing personal identifiers from requests to Large Language Models (LLM). The actual values are then included in the response only on FSM’s end.

  4. Enabling and Using AI Features: We went over the steps required to activate and use AI features, which include meeting certain prerequisites and following specific procedures. The AI functionalities like Assignment Duration Prediction, Intelligent Filtering and Equipment and Activity Summary enhance the scheduling and planning significantly, providing more accurate and insightful information.

Wrapping up, we understood that leveraging AI in FSM further improves dispatching and other field service operations. However, it also requires careful handling to maintain data privacy.

Log in to track your progress & complete quizzes

Learning

SAP FacebookSAP YoutubeSAP LinkedIn