Industrial AI versus ERP AI

We are creating the industrial counterpart to ChatGPT

The criticism of CEO Christian Klein from the SAP community has not been taken out of thin air: The comparison makes one sure! Siemens CEO Roland Busch wants to create 400 billion euros in market capitalization and the industry's ChatGPT, is in the Manager Magazine to read.

The Manager Magazin interview was conducted by Franz Anko-Hubik and Michael Freitag. Roland Busch explained: „Nobody can avoid the topic of industrial AI. It is becoming more and more evident. And we are leading the way in industrial AI. [...] Creating an email with the help of AI is easy. Generating added value in a factory with it, on the other hand, is extremely difficult.“ Obviously, the CEO of Siemens has a unique selling point with AI. But Roland Busch also has a stock market story to tell.

The CEO of SAP, Christian Klein, lacks both: he has no stock market story and currently has no answer to the provocative question of whether AI will eat away at the traditional software of algorithms and data structures. However, it has been clear for twelve months that AI will not only be helpful in existing ERP systems, but will also reshape the system architecture itself.

SAP Hana co-inventor Professor Alexander Zeier is proving this with his start-up Nova Intelligence, in which SAP is also involved. Christian Klein could therefore provide a stringent answer to the threat posed by AI to traditional ERP systems. This important answer will be presented at the Competence Center Summit 2026 of the SAP community on 10 and 11 June in Salzburg. Nova Intelligence Chief Scientist Zeier will be presenting the AI system on the topic of S/4 conversion and SAP Clean Core in Salzburg together with his colleague Nova Intelligence CEO Emma Qian. This keynote in Salzburg is unique in Europe and will be accompanied by analysts from PAC in Munich. Here tickets can be ordered exclusively.

However, SAP CEO Klein is not only lacking in AI storytelling, but also in a clear message to the stock market. You can read the difference in the Manager Magazin interview. Siemens CEO Roland Busch says confidently: „We definitely want to make it into the league of the 50 most valuable companies in the world, or even into the top 30. But I'm not dreaming of Siemens suddenly being worth a trillion euros.“ And where does Christian Klein want to go?

SAP is in a veritable dilemma: the transformation from a very successful ERP system to a new IT technology including the SAP Hana database is proceeding slowly and with many construction sites, see Clean Core and the response from Professor Alexander Zeier. At the same time, the stock market seems to have lost confidence in a stringent AI story at SAP. Of course, SAP also has many and very useful AI tools, but a strategic announcement, such as the one in the Manager Magazin interview with Siemens CEO Roland Busch, is missing.

S/4 transformation without euphoria and vision

For many existing SAP customers, the switch from the tried-and-tested SAP Business Suite 7 (ECC 6.0) to S/4 Hana is less of a voluntary innovation journey than a forced march due to the end of maintenance in 2027, which is technically complex and often difficult to justify from a business perspective. While SAP promotes the new ERP system as the digital core for the intelligent company, in reality the migration often turns out to be a new implementation that ties up immense resources without delivering immediate functional added value.

The obligation to use the Hana database creates a technical lock-in that ends freedom of choice in terms of infrastructure. Even more critical is the shift to the Rise with SAP operating model. Here, customers are forced out of their perpetual on-prem licenses, which have the character of assets, and into a rental model (subscription), which can lead to a loss of autonomy and rising costs in the long term. The threat from SAP management to reserve innovations such as AI functions or the Green Ledger exclusively for cloud customers is massively increasing the pressure on the installation base, although many companies prefer hybrid scenarios.

A future ERP architecture model

In this new architectural structure, the Business Technology Platform (SAP BTP) acts as an indispensable but costly extension layer. As the S/4 core must remain modification-free in line with the clean core strategy in order to be upgradeable, customer-specific extensions are outsourced to the BTP (side-by-side extensibility). The BTP thus becomes the technological toll station for every form of individualization. In addition, the SAP Business Data Cloud (BDC) is intended to solve the historical data chaos through a semantic layer and the integration of partners such as Databricks, but this can also be seen as an admission that SAP has lost data sovereignty single-handedly.

SAP ERP story: Brownfield cemented, greenfield destroyed

However, the biggest hurdle for existing customers remains dealing with decades of custom code (Z developments) that is deeply rooted in the ECC core and massively violates the clean core specifications. Traditional approaches such as brownfield (technical conversion with the legacy code carried over) or greenfield (complete new build) are often unsatisfactory: brownfield cements technical debt, greenfield destroys valuable process knowledge.

This is precisely where the research of Professor Alexander Zeier, the co-inventor of the Hana database, comes in with Nova Intelligence. Zeier, who has first-hand knowledge of the architectural consequences of in-memory technology, has developed Nova, an agentic AI platform that solves the dilemma of S/4 migration through artificial intelligence.

Syntactic and semantic ERP responses

The AI system from Nova Intelligence goes far beyond mere code analyses and clearly shows what SAP is currently lacking or where Christian Klein's storytelling is failing: the AI agents not only scan the old Z-code syntactically, but also extract the underlying business logic (business intent) and document it at specification level. This enables a well-founded fit-to-standard analysis for the first time: the AI checks whether the old in-house development can be replaced by the modern S/4 standard.

If differentiation is still necessary, the Nova AI agents automatically generate new, clean-core-compliant code using AI build - whether as an on-stack extension (embedded steampunk) or as a side-by-side application on the BTP. In this way, Nova Intelligence does not blindly transform the historical legacy, but transfers it into a modern architecture that keeps the S/4 core clean and restores the ability to innovate. For existing SAP customers, this means the only viable way out of the modification trap: an intelligent transformation that reduces technical debt instead of just moving it to the cloud.