Summary
- DSAG says SAP's commercial arrangements for accessing Joule from existing ERP systems could be uneconomic for smaller customers.
- SAP says customers with eligible on-premises ECC and S/4HANA systems can access supported AI capabilities when commercial and technical conditions are met.
- The disagreement centres on RISE with SAP commitments, not a blanket technical prohibition against customers operating a single ERP installation.
German-speaking enterprise software customers are challenging the commercial conditions attached to SAP‘s artificial intelligence tools, warning that smaller organisations could struggle to obtain access without committing to a wider cloud migration.
Thomas Henzler, chairman of the German-speaking SAP user group DSAG, has raised concerns about the terms governing access to Joule from existing enterprise resource planning systems. In remarks to The Register on 18 November, he argued that the commercial model favours large customers able to move individual ERP systems to the cloud while retaining others on their existing infrastructure.
SAP disputes that interpretation. The company says customers operating a single ERP installation can qualify for supported AI functionality if they meet the relevant commercial and technical requirements. It also maintains that customers can make a contractual commitment to cloud ERP before completing the operational migration of their systems.
The disagreement concerns organisations that have invested substantially in SAP ECC or SAP S/4HANA and want to introduce newer AI capabilities without immediately replacing their current deployment arrangements. Both sides recognise the potential value of the technology, but differ over whether the conditions provide a workable route for customers with more limited migration options.
Cloud commitments complicate existing ERP plans
SAP’s approach to enterprise AI has developed alongside its efforts to move customers towards cloud-based software. The company has increasingly concentrated new capabilities within its cloud portfolio, linking product development with commercial programmes such as RISE with SAP.
Many customers, however, continue to operate complex ERP installations on their own infrastructure or through hosting arrangements outside RISE. Those systems can support finance, purchasing, manufacturing, inventory management and other processes that would be difficult to interrupt or replace quickly.
A migration may involve reviewing custom applications, restructuring integrations, testing business processes and changing the responsibilities of internal technology teams. Organisations must also decide when to retire older software, which functions to retain and how to preserve access to historical business records.
In May 2026, SAP widened its approach to AI access by allowing qualifying customers running eligible existing ERP systems to use supported Joule functionality while progressing towards cloud ERP. The concession created a potential route for businesses whose operational migrations would take longer than the development of SAP’s newer AI services.
Henzler’s concern is that access remains tied to a substantial RISE commitment and a minimum relationship between cloud subscription expenditure and existing maintenance spending. He argues that a customer with several ERP systems may be able to meet those requirements by migrating one installation and connecting the others.
A smaller organisation running a single system could face a more difficult financial calculation. According to the DSAG chairman, satisfying the required cloud commitment might involve converting a significant proportion of existing support expenditure into a subscription before the customer is ready to complete its migration.
The financial thresholds have not been disclosed in enough detail to calculate their effect on individual businesses. The available reporting therefore does not establish that every smaller customer will be excluded or that all organisations operating one ERP system face identical conditions.
SAP separates commitment from migration
SAP’s response draws an important distinction between agreeing to a cloud commercial arrangement and moving the systems themselves. The company says eligible ECC and on-premises S/4HANA customers can access supported AI capabilities during their transition, provided they satisfy the applicable requirements.
It also states that eligibility is not determined by the number of ERP instances. A single-system customer may therefore qualify, contrary to an interpretation that would automatically reserve the arrangement for organisations operating several installations.
The disagreement is consequently about the commercial practicality of the route being offered rather than whether the product architecture categorically excludes certain customers. An organisation might qualify technically and contractually while deciding that the associated commitments are too expensive or premature.
This distinction affects how businesses evaluate investment in AI. The cost of connecting an existing ERP system to an automation service may be modest compared with the financial and operational consequences of changing the wider software contract supporting that system.
A company could also have legitimate reasons to retain its present environment temporarily, including manufacturing requirements, regulatory obligations, specialist integrations or the need to complete other technology programmes first. Access to a new AI capability does not eliminate those dependencies.
For customers considering Joule, the decision involves both immediate functionality and the obligations created by the associated cloud arrangements. Without clear terms, businesses may struggle to compare the expected benefits of automation with the full cost of making those commitments.
AI adds pressure to established migration schedules
Joule Studio 2.0 is designed to support the development and deployment of AI agents that work with enterprise applications and business processes. These agents can potentially retrieve information, analyse transactions and assist with tasks extending across several connected systems.
The usefulness of such tools depends heavily on access to reliable business information. An agent working with purchasing data, for example, may need to interpret supplier records, understand approval rules and interact with the applications responsible for financial commitments.
Connecting that functionality to an existing ERP environment can offer an alternative to replacing the underlying system before experimenting with automation. However, an interface that provides access to business data does not automatically establish that the agent will operate accurately or remain within its assigned permissions.
Organisations must still examine authentication, data governance, testing and the responsibilities of employees approving consequential actions. They also need to distinguish between capabilities available today and further developments presented as part of SAP’s longer-term product strategy.
DSAG has acknowledged progress in SAP’s AI tooling while pressing for conditions that accommodate customers with different commercial circumstances. That position reflects a broader concern among established ERP users: technology modernisation can create genuine opportunities without making every proposed change to the underlying commercial model attractive.
The debate is separate from the questions Techopia previously examined about the origins of AI models embedded in SAP software. Model provenance concerns governance and technological dependence, while the current dispute centres on what customers must purchase to use supported AI functionality.
Contract clarity will determine adoption
Large ERP customers often plan system changes over several years because the software supports activities that cannot tolerate prolonged disruption. Cloud migration schedules are consequently shaped by operational readiness as well as the availability of new products.
Connecting AI access to future subscription commitments changes the economics of those plans. Businesses may decide that new functionality justifies the additional expenditure, but that assessment requires reliable information about eligibility, pricing, supported applications and the contractual obligations involved.
SAP and DSAG remain in discussions over the conditions. The company has confirmed that single-instance customers are not automatically excluded, while the user group continues to argue that the existing commercial structure may not provide a realistic option for many smaller businesses.
Until the terms are clearer, customers evaluating Joule alongside established ERP environments will need to determine whether they can obtain the functionality they require without committing to a cloud migration timetable that conflicts with their operational plans.












