Introduction

In this blog post, I am going to share my experience and learning with activating standard Intelligent RPA that came out of the box with SAP, however it did not work as documented. The “payment advice extraction” machine learning model in the Fiori App “Manage Payment Advices” needed a bit of a tweak! Read on to find out more…

Main Content

The Fiori App “Manage Payment Advices” contains an “Import” functionality to upload the payment advice PDF files. If the ML model is not activated, the “Import” is not shown on the toolbar of the App. Please see below how it looks before activation. I have marked a small box in RED color where the button is expected.

Now the required ML model is a standard model on Cloud Platform only. The model usage needs to be registered on your on-premise system using its Service Key ( which is generated after creating a service instance for it ).

The challenge is to activate the machine learning capability. After reading through the documentation to achieve this, we realized that we need to have a ML service called as “payment-advice-extraction”.  To our surprise, though we have “Cash Application” provisioned by SAP, the “payment-advice-extraction” ML model wasn’t available on our global account in Cloud platform. We were certain there is something missing.

So, we started to read through different documentations and we stumbled upon the solution. Actually, as the documentation states that there is a different service, but in reality, it is quite misleading. We did not spot a different service at all. After studying the cash application service key, we saw various service API served within the same key. It looks like this: –

	"api_version": "/api/v1",
	"service_api": {
		"open_items": "/open-items",
		"inference_job": "/inference/job",
		"inference_results": "/inference/results",
		"training_data": "/training-data",
		"training_job": "/training",
		"training_status": "/training/status",
		"payment-advices": "/pa/v1/documents"
	},

Kindly look at the last service API named “payment-advices”. This sounded more like what we were looking for and that was our clue to try same service key instead of looking for a service as per the documentation given.

Wow ! It worked. and we have the IMPORT button on the Fiori App.

Conclusion

The standard bot for “Manage Payment Advices” , requires “Cash Application” which is an additional licensed product and it needs to be configured for the “payment-advices” service API.

Secondly “payment advice extraction” ML model is NOT associated with a different ML on Cloud Platform as documentation suggests. Instead, please use the same key that you use for the “Cash Application” and it should work fine for you.

Thanks ! 🙂