On the 23rd, SAP Japan Co., Ltd. explained its product strategy regarding SAP Business AI.
SAP Japan has indicated a policy of incorporating AI into all SAP cloud solutions, allowing SAP Business AI to work with various SAP applications.
SAP Business AI consists of the AI copilot “Joule” that plays the role of a digital assistant, the “embedded AI function” that incorporates generation AI functions into SaaS, and is customized by users and partners. He emphasized that the company is developing its AI strategy through four layers: the AI Foundation, which can develop its own engines, and the collaboration partner in providing AI functions through collaboration with partner companies.
SAP Business AI overall image
Specifically, Joule, released in September 2023, is embedded across SAP’s cloud enterprise portfolio and provides proactive, contextualized insights from the broad SAP solution portfolio and third-party sources. Data from multiple systems can be quickly categorized, contextualized, and presented with smarter insights. It enables businesses to perform tasks faster and in a secure and compliant manner, helping to generate better business outcomes.
It has already been provided with SAP SuccessFactors and SAP Start, and in SAP S/4HANA Cloud Public Edition, the function is now available for the EAC (Early Adopter Care) program.
Joule architecture overview
Mr. Kiyoshi Iwabuchi, General Manager of SAP Japan’s Business Technology Platform Division, said, “Joule’s architecture loads SAP’s business-related context, data, and system information and links them to create more appropriate solutions based on user input. Joule is powered by SAP’s application know-how.”
Mr. Kiyoshi Iwabuchi, General Manager of SAP Japan Business Technology Platform Division
In addition, Susumu Honna, SAP Business AI Lead, SAP Japan Customer Advisory Headquarters, said, “Joule will greatly change the way SAP is used.The previous UI felt outdated and difficult to use. However, Joule makes it possible to make inquiries and give instructions using natural language. Joule allows for a completely different operating environment than previous SAP systems, and it truly plays the role of a user assistant. In addition, it is possible to use a more appropriate generative AI engine (LLM) for each scenario, and it can also be seamlessly connected to back-end SAP applications through business connectors.
Susumu Honna, SAP Business AI Lead, Customer Advisory Headquarters, SAP Japan
In a demonstration held during the information session, Joule asked the user, “I would like you to tell me three projects that I should take action on,” and the system extracts, displays, and analyzes the corresponding items. We demonstrated that summary information can be easily shared with other employees.
The Joule roadmap, which has already been published, is scheduled to support SAP S/4HANA Cloud private edition, SAP Integrated Business Planning, and SAP Integration Suite after April 2024, and in the future, support for SAP Ariba and SAP It will also be available for Concur, SAP Signavio, SAP Build Apps, SAP Build Process Automation, SAP Analytics Cloud, and more.
Joule Support Roadmap
On the other hand, the number of scenarios in SAP Business AI as a whole, including Joule, is currently over 80, and Joule in SAP SuccessFactors introduced that there are already about 30 scenarios available.
The future roadmap for SAP Business AI as a whole indicates that it will be integrated into various applications and perform various tasks, and says, “We plan to release 49 new scenarios in 2024 alone. “(Mr. Iwabuchi of SAP Japan).
In addition, the AI Foundation has a Generative AI Hub that allows you to select and utilize various LLMs. It also showed that users can use the LLM of their choice with peace of mind under the contract with SAP.
As an AI initiative for SAP Business Technology Platform (SAP BTP), SAP Analytics Cloud Just Ask was released in February 2024. By asking questions in natural language during analysis, it provides functions such as answering in text and automatically generating graphs and charts, and provides the analysis results. In the future, the company plans to provide direct access to models within SAP S/4HANA.
SAP HANA Cloud Vector Engine, released in March 2024, stores vectorized unstructured data and uses it in similarity searches and incorporates context information that general-purpose LLM does not have, thereby improving the answers of generative AI. It is said that it can contribute to improving quality and reliability.
In addition, SAP Build Code automatically generates Java Script and can quickly build data models and business logic for custom apps. Since sample data for testing, which is time-consuming, can be automatically generated, the time required to execute unit tests is significantly reduced. It is suitable not only for business users but also for developers, AI engineers, and data scientists.
SAP BTP x AI: Latest release information
Regarding future functional enhancements, he said, “SAP BTP will release many functions for development sites and project sites.” Joule for SAP Build Process Automation, which automatically generates an ontology representing the SAP Datasphere knowledge graph, and Joule that searches for and automatically suggests appropriate content from predefined workflow and RPA content when you consult Joule with your automation requirements and scenarios. He also mentioned that he offers Joule for SAP Analytics Cloud, which allows you to create charts and dashboards, and perform analyzes and simulations using only natural language instructions.
SAP BTP x AI：Future Roadmap
In addition, more than 30,000 people have taken generative AI-related courses offered by Open SAP, and more than 4.2 million recommended learnings have been proposed per month by the AI embedded in SAP SuccessFactors Learning. It also showed that.
A recent big topic is the collaboration with NVIDIA announced in March 2024. In addition to providing RAG functions jointly developed by both companies, the two companies will collaborate in adding and enhancing generative AI scenarios to be incorporated into SAP S/4HANA Cloud, SAP SuccessFactors, and SAP Signavio. Access federated machine learning (FedML) through SAP Datasphere, accelerate machine learning workloads, and develop generative AI functions and apps with SAP BTP using LLM’s fine-tuning capabilities using NVIDIA’s AI Foundry Services. provide.
Collaboration with NVIDIA
Future initiatives for SAP Business AI in Japan include “deployment centered on SAP’s existing partners,” “support for existing customers and presentation of specific use cases,” and “initiatives that go beyond product boundaries.” ”3 points were raised.
Mr. Honna of SAP Japan said, “SAP partners are not only proposing solutions that incorporate AI, but have also started incorporating AI into their own project management. This year, the BTP Hackathon event is focusing on Business AI, with more than 320 people from 43 companies scheduled to participate, and AI-ready partners will also be launched in Japan. “We have opened AppHaus in Oita City. We will implement prototypes and projects using SAP BTP and work on creating use cases using AI.”