Innovation and digitalization

Generative AI in SAP Products

Where does SAP stand with generative AI?

Since the introduction of ChatGPT at the end of 2022, the topic of artificial intelligence (AI) is no longer only present in the IT industry, but also in the general public.
AI is also becoming increasingly important in the world of work and business, as companies and employees recognize the opportunities for increased efficiency and innovation that it can or could offer in the future.

Since our overview of Joule and AI in SAP products some time has passed. It is therefore time for an update on the possibilities for using AI in SAP applications.

What is AI and what does machine learning mean?

Artificial intelligence (AI) today usually refers to generative AI, which, unlike traditional programs, does not require explicit rules or solution paths to solve problems. Instead, it can solve tasks independently or with minimal input, or develop its own solutions.

The process of building the necessary model is known as machine learning.

There are various approaches and forms of machine learning:

Grafik: Überblick maschinelles Lernen

With the help of self-supervised learning, so-called foundation models are trained on large amounts of data. This turns them into neural networks that can handle a variety of tasks through training and generate output in the form of text, code, images, music or videos.

A special type of foundation models are the so-called Large Language Models (LLM), which are specialized in understanding and generating natural language.

Auch ChatGPT ChatGPT is also based on such an LLM and is therefore able to handle complex language tasks. In companies, ChatGPT is used, for example, for writing emails, writing articles, translating or writing programming code. Prompts are required for this - they are short text entries or instructions that give an LLM the context and direction for generating text. A prompt can contain questions, keywords or sentences that the LLM uses to solve a specific task or generate a desired text.

How can I use AI in my SAP applications?

When using AI, SAP aims to automateprocesses, create a natural interaction for users and provide insights and predictions that complement human decisions.. SAP offers various stand-alone or integrated products for this purpose:

Joule

Joule is an AI co-pilot that can be used across the entire business portfolio. For example, questions about a specific business scenario can be asked in a chat window, which Joule answers by accessing the system data.

Embedded AI Capabilities

Embedded AI Capabilities are AI capabilities that have been “embedded” in existing SAP solutions in order to optimize them.

  • SAP Cash Application

    SAP Cash Application transfers information on incoming payments and open invoices from SAP S/4HANA Cloud to a matching engine that performs a reconciliation and makes assignments.

  • SAP SuccessFactors Human Capital Management (SAP HCM)

    SAP SuccessFactors Human Capital Management (HCM) provides learners with recommendations for content that meets their needs and improves their skills.

  • SAP Concur

    SAP Concur enables the scanning of documents required for travel expense accounting. An AI extracts all relevant header and item data from the document and automatically fills in the required fields in the application

  • SAP Customer Relationship Management (SAP CRM)

    Intrusion detection: In SAP CRM, an AI monitors all access to the system in the background in order to validate, authenticate or, if necessary, block it.

  • SAP Business Technology Platform (SAP BTP)

    SAP BTP Integration Advisor provides more than 10,000 interfaces and scenarios within the Business Technology Platform (BTP).

AI Foundation

The AI Foundation is a comprehensive service that enables AI-supported development on the SAP Business Technology Platform (BTP).
The included Generative AI Hub offers various tools such as the Prompt Editor as an experimentation environment as well as access to various foundation models, with security and data protection guaranteed.

Deep Dive Generative AI Hub

What can the Generative AI Hub be used for?

The Generative AI Hub is included in the AI Launchpad. It can be used to support developments, for example. However, it also offers a wide range of other options.

Overview Generative AI Hub
The Prompt Editor in the Generative AI Hub: an AI playground

The Prompt Editor in the Generative AI Hub acts as a digital playground for experimenting with prompts and models and influencing the result with different settings.

Individual prompts can be given names and assigned to groups (so-called collections) according to topics, similar to ChatGPT. This makes it easy to find the prompts again later in Prompt Management. The result of the prompt can also be saved or recreated by changing the prompt itself or selecting the Foundation Model. Versioning of the results ensures that the individual prompt iterations remain traceable.

In addition, individual parameters and penalties can be used to change and iteratively improve the results.

Bottom Line

It is undisputed that generative AI is already having an impact on the world of work and that this impact will continue to increase. Companies need to prepare accordingly. SAP already recommends activating the existing AI use cases that are available in its SaaS solutions. It also recommends using SAP BTP for customized generative AI extensions.

The Generative AI Hub as part of SAP BTP is an invitation to experiment - companies should start now and find out how they can best integrate AI into their business processes and perhaps even new, AI-driven business models in the future.

Do you have a question or require more information?

Details on the use of personalized data are outlined in our Privacy policy.