Technology Blogs by SAP
Learn how to extend and personalize SAP applications. Follow the SAP technology blog for insights into SAP BTP, ABAP, SAP Analytics Cloud, SAP HANA, and more.
cancel
Showing results for 
Search instead for 
Did you mean: 
Jonathan-Kong
Associate
Associate

JonathanKong_0-1710485238406.jpeg

Motivation

In the fast-paced digital era, businesses seek efficiency in every operation. Yet, the process of creating service orders for faulty equipment remains labor intensive, requiring manual data entry and analysis.

Our solution leverages the power of SAP Build Process Automation as a SAP BTP extension, a web service hosted on SAP BTP Kyma runtime, and generative AI powered by SAP AI Core, to transform this tedious process into a seamless, automated workflow. This blog outlines the approach.

 

The Original Business Process

Traditionally, the creation of service orders is a manual and time-intensive process.  

The cons? Lots of tedious toggling between email and S/4HANA system service order creation page, involving several steps that are not only prone to human error but also lead to delays and potential customer dissatisfaction. 

This is how it originally looks like:

JonathanKong_0-1710483749880.png

Image 1: Outline of the original service order creation workflow

 

  1. Receive customer email requesting a service order for faulty equipment.
  2. Manually extract key information from the email, such as equipment details, purchase date, and analyze attached images for consistency with the equipment description.
  3. Download, rename, and document descriptions of attached images for organization.
  4. Develop a comprehensive problem description based on the email and images, manually entering this along with all extracted information into the S/4HANA system.
  5. Manually upload images to the service order form.
  6. Create the service order.
  7. Confirm service order creation with the customer via email.

 

Solution

We aim to automate the manual aspects of this process, creating a system that automatically processes customer emails, including attachments, to generate service orders in S/4HANA without manual input.

This solution relies on two main components:

  1. SAP Build Process Automation, an extension of SAP BTP, for workflow automation.
  2. Generative AI, powered by SAP AI Core and delivered through a web service on SAP Kyma runtime, to interpret and process data.
     
    JonathanKong_2-1710484384063.png

Image 2: SAP Build Process Automation as an SAP BTP extension

SAP Build Process Automation, an SAP BTP extension, streamlines business process automation and decision-making with minimal coding. It features intuitive drag-and-drop interfaces for easy process design and quick integration within the SAP ecosystem, thereby boosting operational efficiency and facilitating digital transformation.

The automation workflow, created with SAP Build Process Automation, monitors the inbox for service order requests, extracting content and attachments from emails. This information is processed by our web service on SAP Kyma runtime, where the “magic” happens.

JonathanKong_3-1710484533019.pngImage 3: SAP Build Process Automation workflow

Generative AI is an exceptional component in solutions like ours. Within our web service, we utilized the open-source vision model Llava 1.6 to interpret attached images and GPT-4 for analyzing email content, all powered by the SAP AI Core.

JonathanKong_4-1710484594550.pngImage 4: Vision and text models, powered by SAP AI Core

We then use the results to populate the S/4HANA system via POST requests, ensuring that the entire service order form is filled accurately and thoroughly. Multiple image attachments are also supported, with the attached images being uploaded and renamed automatically.

Last but not least, to end off the automation workflow, we can send an automated email reply to the customer with an update on their server order number and status.

 

Results

From just a single customer email,

JonathanKong_5-1710484655981.pngImage 5: Original email sent by the customer

we unlock all these capabilities shown below, hands-free and without any human intervention!😉

JonathanKong_6-1710484715002.pngImage 6:  The service order is successfully created in the S/4HANA system

 

JonathanKong_7-1710484751787.png

Image 7: Results generated by LLaVA 1.6 and gpt-4-32k

 

JonathanKong_8-1710484788624.png
Image 8: Uploaded attachments of renamed images

 

JonathanKong_9-1710484891418.png
Image 9: Automated email response sent to the customer

 

Conclusion

The most remarkable aspect of this solution is its complete automation, which can lead to the significant advantages below:

  • Faster creation of service orders (~2 min per email)
  • Increased accuracy of service orders
  • Higher scalability where hundreds of requests can be handled simultaneously
  • Better customer satisfaction due to rapid response and accuracy

Beyond service order creation, the potential applications of AI and SAP products can extend to various other business processes, including warranty claim automation, invoice posting automation, and many others.

This video summarises the automatic service order creation process using SAP Build Process Automation and Generative AI:


If you’re interested in the technical details of this solution, you can visit this blogAlso, if you're interested in exploring solutions like this or wish to learn more, feel free to reach out to me.

 

 

References

SAP Build Process Automation
SAP AI Core
SAP AI Core service guide
Generative AI @ SAP
Service Order documentation
Generative AI hub sdk

 

 

2 Comments