During the Sapphire conference earlier this year, we noticed that SAP has taken much more control of the organization and its developments. The company is also making better choices, something that hasn’t always happened in the past, according to SAP CEO Christian Klein. SAP is steadily extending its portfolio and taking the right steps in AI development.
Last month, the virtual edition of SAP TechEd took place. It is SAP’s developer conference, which familiarizes developers with the SAP platform, the latest developments, and new products. For us, it is a good time to see how far SAP has come with all the SAP Sapphire announcements earlier this year, especially following the interview with Christian Klein.
It has not escaped our attention that the stable SAP of six months ago has its challenges. Recently, several people left the board. There were also announcements during SAP Sapphire that are not available to this day, even though they were supposed to be. That’s unfortunate, though.
Fully automated AI migration to SAP S/4HANA is delayed
Most notable in this case is the almost fully automatic migration from SAP ECC (the outdated SAP ERP system), to the new SAP S/4HANA. The customized source code in ECC and the data stored in the solution should be migrated more easily through AI. The source code is scanned, and the customized code is converted into plug-ins or extensions for S/4HANA. Data in the old system can be migrated and enriched with AI, using, for example, product flyers, specification sheets, or marketing material to act as grounding (source material) to enrich that data. During Sapphire, SAP made it appear that customers could use this immediately. Now it seems that this is still some time away.
We talked to Jesper Schleimann, SAP’s head of AI. He revealed that while these AI functions are already available internally at SAP, external SAP consultants will have to wait until the first quarter of 2025.
Many investments in AI
SAP is heavily invested in AI, just like many other SaaS platforms. Both in predictive AI and generative AI. During TechEd, new Joule assistants will also become available within various SAP solutions, allowing customers to do more and more with the internal chatbot.
However, we also see some players in the market already moving on to next-generation AI, in the form of AI agents. SAP is also working on AI agents and has released two AI Agents. An AI Agent for Automatic incident handling, and for finance, specific processes are made more intelligent with an AI agent. Think about processing invoices and paying them.
We asked Schleimann what else we can expect from SAP AI Agents. He indicated that these are the only two at the moment. He especially emphasized the complexity of having the proper guardrails in an enterprise environment. SAP customers expect a robust AI framework that provides reliable and explainable answers.
SAP is leading the way in the ERP landscape with its AI agents, but if we take it more broadly, there are already parties a bit further along. Schleimann says he expects that by mid-2025, customers will be able to build AI agents themselves. Customers can then configure and develop AI agents themselves via low-code/no-code.
Knowledge Graph is foundation for SAP AI
At SAP the Knowledge Graph will act as the foundation for their AI and AI agents. During TechEd they announced a massive expansion of their Knowledge Graph. In total, there are now 450,000 ABAP tables, 80,000 views, and thousands of SAP Fiori applications linked to the Knowledge Graph. The SAP Knowledge Graph is linked to the SAP AI Core, which developers can use to develop AI features within the SAP platform.
The big advantage of a Knowledge Graph is that you not only have data at your disposal but also the context and all the associated metadata. This lets you feed an AI much better with how data should be interpreted and read. A very simple example: you can have a score from 1 to 100, but the metadata and context ultimately determine whether 1 is very good or very bad.
Of course, context goes much further. Consider sentiment, emotion, customer relationships, processes, the stage of a process, and its dependencies. That can often all be extracted from context and metadata.
SAP is taking the right steps toward AI
All in all, we can now safely say that SAP is steadily building on all its AI innovations. It also took a nice step during TechEd with all the extensions in Joule. Customers will soon have more options with the chat assistant. SAP is a forerunner in the ERP landscape.
SAP is taking the right steps in terms of development. The data model is much simpler, it invests heavily in the Knowledge Graph, and the first AI agents are ready. Step by step, we see SAP fulfilling its promises. They have to keep that up in the coming period to continue to innovate next year.
New people, same course?
SAP is looking for new executives to fill the gaps in their management. These will have to be experienced top executives if SAP wants to continue to innovate at this pace in the future. The coming months will reveal whether this will affect the organization’s strategy. Will the course remain the same, or will these new executives come with strategy changes?
Automated ERP migration can make a big difference
Finally, we are very curious to know when customers can migrate their legacy ERP environments to S/4HANA using AI in a genuinely automated way. We think that will make a big difference for customers and SAP itself. Using AI to read out all that custom code and convert it to extensions for S/4HANA will save enormous time. The same goes for reading out and enriching data sets. That will save a lot of time and, therefore, costs. If that is widely available, it can also benefit SAP’s results. Many customers are still running ECC migration projects. However, the question is whether SAP partners will also be eager to offer this automated migration. After all, the cost and time savings also mean that partners can write fewer hours.