Cleanfield Strategy and SAP Clean Core
The Clean Core strategy championed by SAP CEO Christian Klein and his executive team is far more than just a technical update. Clean Core marks a radical paradigm shift (Cleanfield)—one that is painful for many existing customers—that finally brings an end to the era of architectural freedom seen in the R/3 and ECC eras. In the past, the ABAP programming language was used to tailor standard software precisely to individual needs through millions of lines of Z-programs and modifications—a “toxic legacy” that is now hindering modernization efforts.
A technical analysis of the global SAP community makes it clear that an existing SAP customer should not implement the Clean-Core program not for aesthetic reasons, but because it is the fundamental safeguard for the economically viable use of artificial intelligence, for the urgently needed ability to upgrade in the cloud, and, in the long term, the path to the Autonomous Enterprise.
SAP BTP as a Strategic Anchor
The central architectural guiding principle is: „Keep the Core Clean.“ The digital core of S/4 must remain completely untouched, while all customer-specific extensions are consistently decoupled and migrated to external platforms. The Business Technology Platform (SAP BTP) serves as the indispensable technical backbone and the “playground” for innovation. As a Platform-as-a-Service (PaaS), it provides the necessary framework for side-by-side extensions that communicate with the core system via standardized APIs (Note: SAP API Policy) without compromising its stability.
As part of this strategy, existing SAP customers must rely on modern development tools such as ABAP Cloud, the ABAP RESTful Application Programming Model (RAP), and the SAP Cloud Application Programming Model (CAP). These models enforce compliance with Clean Core rules through their compilers, as they technically prevent access to unapproved SAP objects. For business units, the low-code/no-code platform SAP Build is also gaining importance as a way to implement simple customizations quickly and in line with standards.
Cleanfield: The Smart Third Way to Transformation
In the debate over migration paths such as Brownfield (technical conversion, including all legacy issues) and Greenfield (new implementation with loss of process knowledge), Cleanfield has emerged as an innovative third option. This approach, made possible largely by the startup Nova Intelligence and its Agentic AI platform, relies on an „archaeological“ analysis rather than mere translation.
Autonomous AI agents semantically analyze the entire historical code base and reconstruct the actual business logic (business intent). In a systematic “fit-to-standard” comparison, the AI identifies which former custom solutions can now be covered by the S/4 standard. Only the value-adding differentiators are then regenerated for the target architecture in accordance with the Clean Core standard. Cleanfield thus promises a lean S/4 system that is upgradeable from day one without sacrificing business-critical know-how.
In the current S/4 transformation landscape, an interesting rivalry has emerged between two technical powerhouses, both of which promise to cut through the Gordian knot of legacy custom code and technical debt through the use of Agentic AI: Nova Intelligence and Conduct.
Lift-and-Shift vs. Agentic AI
For existing SAP customers faced with the choice between a risky technical „lift-and-shift“ (brownfield) or a costly fresh start on a greenfield site, Nova and Conduct mark a paradigm shift away from manual consulting work toward algorithmic intelligence. While Nova Intelligence is deeply rooted in Walldorf’s DNA and shaped by SAP HANA co-inventor Professor Alexander Zeier as well as AI expertise from Google DeepMind and Meta AI, Conduct stems from the team of data specialists at Palantir Technologies.
SAP HANA, Clean Core, and ABAP Cloud
With its proprietary Cleanfield approach, Nova positions itself as a sort of “archaeological architect” of the ERP system. The system uses autonomous AI agents not only to translate the accumulated ABAP code syntactically, but also to reconstruct its business logic—known as “business intent”—and document it at the requirements specification level. A key architectural differentiator for Nova is its native access to SAP-internal primitives such as the transport system, the indexing of the entire code base, and direct integration with modern platforms such as ABAP Cloud or RAP.
Nova is explicitly designed as a lifecycle-oriented platform intended to keep the S/4 system close to the standard and „clean“ even after migration. Reference customers such as Festo and the Kion Group report dramatic efficiency gains, with code analyses that used to take weeks now being completed in just a few days.
AI Operating System for the Autonomous Enterprise
Conduct.ai, on the other hand, positions itself as the „AI operating system for enterprise software.“ Founded by former Palantir engineers Jan Philipp Haas, Philipp Höfer, and Henry Thompson, the company leverages its expertise in handling massive amounts of data and complex scenarios to help IT teams transform their ERP landscapes. Similar to Nova, Conduct relies on agents that analyze legacy systems, generate documentation, and automatically migrate custom code into the S/4 environment.
A key indicator of Conduct’s strategic importance is the fact that SAP is not only acting as a partner but has also invested directly, through its investment arm, in the London-based startup, which has already raised $60 million in venture capital. Conduct is already being used by industry heavyweights such as Daimler Truck, DHL, and Heidelberg Materials to bridge the cognitive gap between outdated processes and the new digital reality.
A direct comparison reveals a difference in strategic focus: While Nova draws its strength from deep vertical integration and an understanding of SAP’s DNA as well as specific Clean Core requirements, Conduct excels with a more horizontal approach centered on an overarching IT governance system. Nova acts as a specialized intelligence layer that integrates seamlessly into SAP governance and supports developers directly at the code level. Conduct, on the other hand, leverages Palantir’s methodological rigor and “intelligence-first” mindset to orchestrate the transformation as a whole.
The History of R/3 and the Future of S/4
Existing SAP customers are currently trapped in an architectural straitjacket that SAP’s strategists are promoting as an unavoidable development doctrine under the banner of “Clean Core.” Viewed critically, this concept means that a company’s S/4 digital core must remain completely untouched—which, for many organizations with millions of lines of historically accumulated Z-code, is akin to open-heart surgery.
The goal is a responsive IT architecture that remains low-maintenance through standardization and enables seamless cloud upgrades as well as the integration of AI innovations. For existing customers, this is critical, as technical debt in outdated systems hinders innovation and causes total cost of ownership (TCO) to skyrocket due to time-consuming manual testing with every release change.
SAP Autonomous Enterprise
The era of monolithic SAP modifications is irrevocably coming to an end: Both Nova and Conduct demonstrate that a customer’s technical autonomy today depends on how quickly they can migrate their accumulated in-house developments into a near-standard, AI-enabled architecture. Choosing between the two platforms is a choice between the specialized depth of an SAP insider (Nova) and the methodological breadth of a data intelligence expert (Conduct). Since SAP acts as a sponsor or investor for both providers, it becomes clear that SAP itself has recognized that the path to the Autonomous Enterprise can only be achieved through such intermediary bridges that intelligently dispose of customers’ historical baggage, rather than simply shifting it to the cloud.
In summary, Clean Core is not an end in itself in the IT realm, but rather a prerequisite for transforming the ERP system from a rigid data graveyard into a responsive Autonomous Enterprise that generates real competitive advantages based on clean data and upgrade-stable logic.
SAP LaMa: End of Support in 2027
