The gap between SAP and its customers must not widen further

SAP has taken a beating of late in the stock market due to perceptions that company’s enterprise software offerings and foothold are vulnerable to the rise of AI. Now, SAP customers are voicing their concerns — less about the replaceability of SAP platforms at the hands of AI than in terms of the AI outcomes and clarity they are getting from SAP’s platforms and vision.

At last month’s German-speaking SAP User Group (DSAG) conference, the overall sentiment was clear: There is still a long way to go between SAP’s ambitious AI plans and the reality its customers face. Stefan Nogly, DSAG’s technology expert, warned in an interview with Computerwoche against further divergence — but also says he sees some progress.

“We need to be careful that the gap between SAP and its users doesn’t widen further,” he says — a concern recognized by SAP itself, as SAP CTO Philipp Herzog admitted in his keynote address at the event that a significant gap exists between AI innovation and actual outcomes.

“SAP intends to actively improve in this area. I am generally satisfied with the answers and the announced measures,” Nogly adds.

The top tier: AI for IT

Nogly understands why many companies remain hesitant regarding AI and SAP, as the DSAG Investment Report 2026 recently revealed. Integration of AI agents into business processes is essentially the “final stage” — and in many cases, trust, experience, and, above all, a suitable data foundation are still lacking. “We are in a phase where we have a lot to learn and try out,” said the DSAG spokesperson.

From Nogly’s perspective, it makes sense to promote AI experimentation first within IT. SAP has already announced its intention to provide greater support in this area — for example, through migration tools and additional AI functions within IT and transformation processes. “Often, the initial focus is on coding support, such as through ‘Joule for Developers.’ However, there are actually many more areas of application,” Nogly explains.

Especially in the context of cloud transformations, AI can significantly contribute to efficiency, for example, in adapting interfaces. Many companies have not just a few, but hundreds or even thousands of interfaces — from business-to-business to application-to-application — that need to be adapted and optimized. Here, AI can significantly increase speed and productivity. The same applies to user interface development, says Nogly. If developers can create multiple UI variants more quickly and coordinate them with the relevant departments, the benefits are immediately apparent.

Overall, there is a wide range of potential applications for AI, where the added value often becomes apparent more quickly than with direct integration into business processes.

In search of added AI value

Many companies are not yet ready for such integration, however, the DSAG representative adds. The industry is currently in a learning phase, he says, with the focus primarily on gaining experience and understanding where AI actually delivers added value.

To that end, Nogly recommends testing AI in a controlled manner, with clearly defined areas of application, developed step by step. More complex use cases, such as those SAP is currently strongly promoting, is not yet within reach of many companies, for example, when it comes to public cloud scenarios or the use of data products in the Business Data Cloud, Nogly adds. This level of maturity takes time to build — nevertheless, customer companies expect SAP to demonstrate a clear and practical path to get there.

For that, companies primarily need planning certainty, he says — and time to uplevel operations. “This takes a bit of time, and we need to allow ourselves that time. We should consciously say: We’re trying things out and learning,” a process that also includes fundamental strategic realignment within customer companies themselves.

Some pioneers closely aligned with SAP’s strategy, such as Frosta and Hörmann, have demonstrated that SAP’s approach works in principle — however, such flagship projects, highlighted in the event’s keynote, have been rather isolated. Many midsize companies, in particular, are still acting cautiously, observing costs, benefits, and risks, and waiting to see how things develop.

A new dimension of security

A key issue in this context is security. Nogly emphasizes that, for example, critical infrastructure companies in the energy, transport, and healthcare sectors already have to comply with very strict requirements under the IT Security Act (IT-SiG 2.0) — regardless of AI — while for the wider economy, the requirements of the NIS2 Directive and the BSI IT Baseline Protection serve as the benchmark. “This must become the standard practice for any company,” he says. However, AI adds an additional dimension.

“Many people find it more enjoyable to talk about productivity and simplification,” says Nogly. “But we also need to know precisely what data AI accesses, whether it modifies data, and how decisions are made.”

Trust in AI systems can only be built through security. Therefore, a new discipline is emerging within IT security. Nogly warns against focusing investments solely on efficiency gains: Companies must also invest in understanding the technology and its risks. “Those who only look at productivity and neglect security are missing the mark,” he says.

The situation is becoming increasingly complex, especially with regard to AI agents taking on increasingly autonomous tasks and linking processes together. This development marks a new level of complexity — and significantly increases the demands on governance, control, and security mechanisms.

AI needs data — and patience

A key obstacle for companies remains the data foundation. For many SAP customers, analytics landscapes are fragmented, and a unified data layer is lacking. “This is the reality for the majority of companies,” says Nogly.

With its Business Data Cloud (BDC), SAP has chosen a sound strategic approach, with concepts like data products and a semantic layer fundamentally suitable for bringing order and transparency to data landscapes. But the solution is coming late: Numerous organizations have already invested in platforms such as Snowflake or Databricks to address precisely these problems. Accordingly, the question now arises as to how existing solutions can be meaningfully combined with the BDC — without adding complexity or high costs. “This needs to be explained,” says Nogly. Introducing yet another tool is neither trivial nor inexpensive.

Furthermore, he sees room for improvement in SAP’s implementation: The product needs to mature further, become more understandable, and be more accessible. Besides technological hurdles, the commercial model also plays a role. “The idea is good — but it still needs to be proven,” he summarizes.

Public cloud ERP systems haven’t yet reached sufficient maturity to be a viable alternative for the majority of customers, Nogly points out, though he no longer considers implementation of SAP BTP (Business Process Transfer) to be a major obstacle. The fact that many companies still rely on on-premises or private cloud models is primarily due to the realities of the transformation process: companies have to prioritize. Often, the ERP system migration comes first, followed by facets like analytics or data platforms. “It doesn’t all happen at once,” Nogly emphasizes. Limited resources, budgets, and organizational capacity mean that the transformation can stretch over years.

This context also clarifies why many AI initiatives are still in an early stage. Only when a solid data foundation exists can AI applications be used effectively and scaled.

“We talk a lot about AI these days, but at the same time we’re still in an experimental phase,” says Nogly. The pace of new models and applications is rapid — but their actual implementation in companies is lagging behind.

Pressure on IT is increasing

At the same time, the pressure on IT departments is increasing. Nogly reports that many CIOs are currently being confronted by their management with AI initiatives. “Everywhere, solutions are supposed to be tested quickly,” he says. This approach often contradicts necessary foundations such as data quality, security, and governance, creating a tension between the pressure to innovate and technological reality.

Regarding the question of standardization versus individualization, the DSAG representative also advocates for a clear course. The goal must be to create stable and maintainable systems. “We want to move away from a situation where a single patch terrifies an entire company,” he says.

SAP has laid the right foundations with its Business Technology Platform. However, the platform and its associated extension concepts now need to be understood and consistently used in practice, says Nogly. “First, you have to fully explore its potential and learn how to use it.” This includes technologies such as Fiori and CDS Views, which will play a central role in the future.

SAP has clearly confirmed this direction: “Philipp Herzog said on stage: Absolute investment protection in Fiori, in CDS Views, in this entire underlying framework. Yes, that is the future,” Nogly notes. However, this also means a profound transformation for companies, he adds. Developers must move away from classic ABAP approaches and understand and apply the new platform landscape.

Once this step is completed, further expansions will still be possible — but within clearly defined guidelines. The goal is a platform approach that ensures stability and security and eliminates the fear of updates, according to Nogly.

At the same time, Nogly points out that this change also takes time. SAP began its transformation 10 to 12 years ago, whereas many companies started much later — some seven or eight years ago, others only now. Consequently, their levels of maturity vary considerably. In many organizations, a fundamental rethinking of operations and development is only just beginning.

The necessary technologies and expansion options are fundamentally available. But it remains to be seen whether they will be sufficient in every case. Furthermore, the possibility of integrating other solutions into modular IT landscapes still exists.

DSAG demands more clarity, maturity and support

During its Technology Days event, DSAG also compiled a list of demands, primarily calling on SAP for more clarity, maturity, and support in implementing key future topics.

• For AI to truly become enterprise-ready for SAP customers, orchestrated agents, transparent decision logic, secure data, and open integration for third-party agents are needed. At the same time, DSAG expects a clearer strategic vision, simpler implementation, and investment protection for existing technologies such as Fiori.
• In terms of data, the focus is on expanding Business Data Cloud. It should serve as a unified, trustworthy data layer. This requires clearly defined data products, improved cataloging, and practical migration paths to modern data architectures.
• In ​​security, DSAG seeks binding best practices, clear governance models, and, above all, transparency and traceability of decisions — both technical and regulatory.
• For transformations, the user group would like more concrete support: for example through funding programs, more migration tools, more practical reference architectures, and closer coordination with SAP on roadmaps.