Preamble

Before we start delving into this topic, I would first like to define the scope of what we are talking about when we say AI-assisted use case (you may also find sometimes the terms ‘Embedded AI’, or ‘SAP Managed AI’ being used to describe the same things). Well, put simply, these are AI use cases or scenarios that run through an SAP Managed BTP as a conduit to an AI generative service, that is provisioned, and fully managed by SAP. This in contrast to an AI service that runs through your very own SAP AI core setup in your own BTP. At the time of writing this post, the SAP managed AI services that are candidates for this post are the ones depicted by the image below and accessible by this link to the SAP Discovery Centre (they are found by filtering on the product = SAP S/4HANA Cloud Private and Commercial Type = Premium). By the filter and title of this blog you will have understood that this applies to SAP S/4HANA Cloud Private Edition. These capabilities are not available in SAP S/4HANA On-premise deployments. In addition, typically these AI uses cases require for a commercial construct to be in place, i.e you have purchased AI Units.

The process

The process required to activate an AI-assisted use case, at a high level, has to follow the steps below. We will go into the details of each step:

1. Identify the AI use case(s) that are relevant and applicable to your organization
2. Gather the information that is required throughout the activation process
3. Request your SAP Managed -BTP- AI instance
4. Perform the AI use case setup prerequisites
5. Perform the ISLM connection setup
6. Perform the AI use case activation
7. Perform application specific configurations
8. Test and use the AI scenario

1 Identify relevant and applicable use cases

If you are not sure where to start in your AI Journey, then I would recommend you first start by executing an AI Discovery Workshop to understand what AI is, what it can do for your organization, identify the high value use cases or headaches you want to solve with AI, build a roadmap and then leverage the available SAP delivered AI-assisted use cases (whether base or premium).

If you want to just get going and trying anything to see what sticks, then I would in any case recommend that you make your first point of call the SAP Business AI Discovery Centre AI Catalog, which lists the entirety of SAP’s AI offering. Having applied the required filters, you will then be able to view the offering that applies to SAP S/4HANA Cloud Private Edition, and SAP Managed AI scenarios. For each of the available scenarios, you will find a lot of very valuable information that will help you to understand what the use case is, its benefits, where to start, a business value 3 pager and more. Taking the AI-Assisted Journal Upload scenario as an example of the documentation available for a given scenario is shown below:

One you have taken stock of this page you should be in a position to know if the scenario is relevant and applies to your organization, i.e:

• Am I using the underlying functional area that is required to benefit from this use case? (There is no point trying to activate the AI-Assisted Conversational Planning scenario if you do not use Advanced Transportation Management!)
• Am I technically on the right platform to activate a given scenario? (Note that not all scenarios are available for all versions of SAP S/4HANA, or indeed for all FPS’s of a given version. The general rule being that the more recent (latest and greatest) your platform is, the broader the offering you can access.
• Are there any other prerequisites? This could vary, i.e:
  • Some scenarios are ‘Hybrid’ scenarios and require for Joule to be already fully activated,
  • Some scenarios require for Fiori to be available (either SAP Fiori Launchpad or Workzone) – no scenario runs in the SAP GUI!

2 Gather the information required

If you are moving on to this section, then I assume that you have identified the scenario(s) you want to implement and deemed them as being relevant and applicable!

For each AI use case, you will need certain pieces of information that will be called for at different points of the setup. I recommend that before you proceed on to step 3, you have those pieces of information fully identified and available – this will confirm that you have read the SAP help and not guess your way through the activation!

2.1 The technical name of the AI Scenario

Each scenario (eg: AI-assisted Journal Upload) will have a specific technical name, which is the ISLM (Intelligent Scenario Lifecycle Management) name of the scenario. You will need this technical name when you request the SAP Managed BTP instance and when you configure a scenario in your system. Taking the AI-assisted Journal Upload scenario as an example, we can find this name in the SAP help in the Prerequisites / setting up the system section, as shown below, and we can see is FIN_INTLGNT_JRNLUPLD :

2.2 The OAuth 2.0 Client Profile

The next piece of information you’ll need to collect is the OAuth 2.0 client profile. You will need this when you configure the ISLM scenario in your system. Again staying on the example of the  AI-assisted Journal Upload scenario, we can find this information in the SAP help in the Prerequisites / Setting up the Connection to Intelligent Scenario Lifecycle Management, as shown below and we can see is S4_PR_IJU_FIN :

Note, that not all scenarios will list a OAuth 2.0 Client Profile Name – but you do need this information! In this case, I would suggest you follow the guidance provided in this SAP help page and use one of the profiles mentioned in the section Procedure.

3 Request your SAP Managed BTP -AI- instance

As previously mentioned, these AI use cases will run through an SAP Managed Instance. You must request it using the familiar SAP support process, creating an incident on a specific component. The process that you must follow to request this SAP Managed BTP Instance is detailed in the SAP help page Requesting Access to SAP AI Services on SAP Business Technology Platform (SAP BTP).

Note that you will need to follow this requesting process for every combination of AI Scenario and System ID and Client – do not attempt to re-use previously provided assets. When you create the incident, make sure that you user ID also has the appropriate authorizations to create Private Cloud Service Requests (See SAP Support note 2669783 for more information)

The SAP Support team that will receive your ticket are well versed in handling these requests, and I suggest you follow the below template to expedite your request.

• Message Title: Requesting SAP Managed BTP – AI Instance
• Component to use : CA-S4H-BTPAI
• Priority : As required
• Message Body: As per SAP standard process, herewith this message requesting provisioning of an SAP Managed BTP – AI Instance with the following details:
  • S/4 System ID and client where the ISLM will be configured : ABC / 123 (Your own system ID and client)​
  • Name of the intelligent scenario: FIN_INTLGNT_JRNUPLD (Journal Upload) (The code/name of the Scenario for which you are requesting an SAP Managed BTP)​
  • Preferred Authentication Method : OAuth 2.0 with client credential flow  (preferred) OR OAuth 2.0 with mTLS *​
  • Preferred Data Center : EU10​

* Choose one – if mTLS, then provide a certificate as attachment to your upload.

Text in Orange Colour denotes text that you must amend /personalize as required.

Once SAP Support receives your request, the team will verify that the request can be fulfilled (You have AI units and the system is a RISE system). If yes, once provisioned, they will create a service request for you, that you must submit in order to be able to download from it several digital assets, namely a JSON service Key and one or more SSL Certificates. Download these and store them in a safe and secure place. You will need them in order to securely setup the connection from your SAP S/4HANA System to the SAP Managed BTP instance.

4 Perform the AI use case setup prerequisites

Once your reach this point you will start to configure your system per se and carry out the following prerequisite activities.

4.1 Setting up the ISLM Background Jobs

Several background jobs are required to be setup, so that the ISLM scenario is able to stay updated and pull the required content from the SAP managed BTP. In order to setup the required background jobs, follow the instructions in the SAP Help, indicated in this SAP help page. This is a one time activity – it does not need to be done each and every time you activate a new scenario – only the first time.

4.2 Setting up the Trust to the SAP Managed BTP instance

As previously indicated, SAP support will have provided you with 1 to 4 SSL certificates. You must import these SSL certificates into your system, using the transaction STRUST. Further information regarding where and how these certificates must be imported is detailed in this SAP help page. SAP support will also provide you with a more detailed guide in the service request, when they hand over to you the SSL certificates in case you are new to this and require a bit more guidance to do this.

5 Perform the ISLM connection setup

In this step you will maintain the connection to the SAP Managed instance and it is here that the pieces of information you collected at the beginning (the technical name of the AI (ISLM) scenario, the name of the OAuth 2.0 Client Profile and very importantly the JSON service key that SAP support provided you. The procedure required to perform this step is described in the SAP help here, Maintain Connection for an Intelligent Scenario.

However this step is sufficiently important, that I will add some guiding screenshots below. To help you follow along, we will go through the activation of a Hybrid AI-assisted use case, in the area of Sales and Distribution. The specifics of this feature is document here, and we will be activating the AI-assisted Sales Order Fulfillment scenario.  Make sure to always read the SAP help before you start configuring a scenario. Because the system I use is based on SAP S/4HANA Private Cloud release 2025, there is a subtlety I must be aware of – the scenario and model setup names are different to lower releases!

To recap before I get started the information I will need is:

• The technical name of the scenario : SD_INT_AI (and I will be interested to select the SD_INT_AI_GPT4O model)
• The name of the OAuth 2.0 Client Profile: I could not find it in the help page, thus I will select ISLM_SAPGENAI_OAUTH_PRF_HCP (for Generative AI) that was documented here,  if one is not auto proposed when I do the configuration
• I have the JSON service key ready that was given to me by SAP Support.

You can access the ISLM connection mapping via the transaction code ISLM_CONN_MAP. Start by creating a new connection.

Use the search help to select the scenario for which you want to create a connection.

In the next popup, you will need to (1) enter the OAuth 2.0 Client Profile. If it is filled for you, keep the proposed value. Paste the content of the JSON service key  SAP Support provided to you, as shown in (2), and then click Next.

The connection details will be validated, and you can see that SAP will create certain artifacts for you, namely (1) an RFC destination, and (2) an OAuth 2.0 Client configuration.

You are also (3) invited to check the connection so go ahead and do that by clicking (4) – Connection Check.

What you are looking for is (1) a ‘Green’ Connection status indicator. You can then (2) save the profile. If you get an error message, then please check the section later ‘Troubleshooting common problems’.

Once you return to the main connection screen, you will see your scenario (1) appear in the list. You can also at anytime here perform a connection check (2).

Now that the scenario is saved, your SAP system will start to sync the AI content to your system. Give it 5-10 minutes before moving on to the next step!

6 Perform the AI use case activation

Again referring to the SAP help, the activation process is described here in the page Activate and Deactivate. I will again here provide some screenshots. For this step you need to navigate to your Fiori Launchpad and open the app named ‘Intelligent Scenario Management‘ app.

You can (1) use the available filters to narrow down the list of scenarios and locate (2) the one that you are interested in. Again the devil is in the detail. Note that there are other scenarios named SD_INT_AI*, so I have to make sure to pick the right one ! As per the SAP help, and the scenario for which I created the connection, I select SD_INT_AI. I click on the line to go into the details.

As per the documentation for the scenario, 3 models are available for me to choose from. I will select SD_INT_AI_GPT4O, and then click on the line to progress in its configuration.

A new pane will open to the right, showing me ‘trainings’. I select (1) the training and then click (2) to ‘Deploy’ the model. Note that at this time the number of deployments is 0 (3) – meaning none exist yet.

A popup will appear showing you the (1) model name and version, then click (2) to deploy the model. Note in some cases the model version field could be blank. In those cases I would suggest you enter ‘latest‘.

The deployment will then progress.

And hopefully finish with the status (1) ‘Deployed’ if all goes well. You then need to (2) select the deployment and activate it (3).

Once the activation is complete, you will see this information indicated, as shown below.

The scenario is now ‘live’.

7 Perform application specific configurations

This section will vary for each scenario that you are trying to activate – there is no set rule. Whilst usually there will at minimum be a need to fine tune or create roles and / or authorizations, some scenarios will require you to perform application specific configurations. For example the AI-assisted conversational planning scenario requires that you perform some configurations in the area of your transportation planning layouts, so make sure to read the SAP help or consult the prerequisite SAP support notes that provide you this detail.

8 Test and use the AI scenario

Once your setup and configuration is complete, you can proceed with testing your scenario. In this case we activated a Hybrid (Joule) use case, so let’s try a prompt associated with this capability! Success !

We have now activated an Ai-assisted (Embedded AI) Scenario!

 

Troubleshooting common problems

I am also below documenting some of the more common issues that I have seen implementations come across.

HTTP failure in connection mapping.

When you create a connection mapping, and upon checking the connection, you do not get a ‘Green’ light, but instead an error message saying ‘HTTP failure, processing failed, invalid state, invalid timeout or others‘, and if your system is hosted on a hyperscaler, then you need to perform some additional configuration.

 

So go ahead, save your connection mapping (with an error state), then head on to the SAP help and check this troubleshooting section. Essentially, you will need to maintain a proxy and a port in the RFC destination and OAuth 2.0 client configuration that were auto generated when you created the connection mapping. So go ahead and maintain the values ‘proxy’ and port ‘3128’ as shown below.

In the ‘Technical Settings‘ tab of the RFC destination:

And in the Access settings of the OAuth 2.0 client configuration:

If you are unsure about the names of the RFC destination or OAuth 2.0 client configuration, you will find those in the connection mapping settings, for your AI scenario.

Once you have made those changes, don’t forget to go back to the ISLM connection mapping transaction (ISLM_CONN_MAP), select the faulty connection and press the ‘Check Connection‘ button. It should now display a Green – OK – indicator.

Scenario is not registered or published

If you try to create a connection mapping for a  scenario, and you get an error message such as ‘The scenario <NAME_OF_THE _SCENARIO> is not registered or published’, and if you are on SAP S/4HANA 2023 FPS01, then please raise a Service Request with your nominated ECS contact and ask them to verify SAP support note https://me.sap.com/notes/3437407 – they will know what to do.

Cannot Deploy and activate a scenario

You have done some configuration that you believe complete, but when you come to deploying and activating the scenario, you find that you are not able to do so, and in addition, you see a message box like below in the ‘Intelligent Scenario Management‘ app.

Assuming you have been able to create the connection mapping without any errors, the likely culprit is that you have not set up the ISLM background jobs. Check that all required ISLM jobs as described above have been setup.