Agentic AI for SAP

With Mastering SAP Collaborate, an SAP TechEd on Tour event, arriving at ICC Sydney on 18–19 November, SAPinsider has been speaking with sponsors about what makes the SAP market in Australia and New Zealand distinctive. The most recent conversation was with Joel Stehr, Go-to-Market Specialist, SAP on AWS, at Amazon. Stehr sits within AWS’s ANZ team, where he leads go-to-market and partner development, including management of the SAP alliance.

The landscape that Stehr described is that the question for many Australian SAP customers is not whether to move to the cloud. It is how to turn data already sitting in the cloud into value, and how quickly they can put agentic AI for SAP to work in the near future so that they don’t fall behind competitors and others taking more aggressive approaches implementing AI.

A Cloud-Mature Market Asking “What’s Next?”

According to Stehr, SAP customers in ANZ are more mature in their approach to cloud than peers elsewhere, and it is a difference that stands out when he compares notes with colleagues across Asia Pacific and Japan. “Most of our conversations are not with customers who have a legacy on-premise environment they’re looking to modernize,” he explained. “Most customers are saying that is a journey they started years ago.”

That maturity helps explain what has been communicated by other event sponsors. Some describe Australian organizations as lagging on the move to SAP S/4HANA, held back by budgets and a need to show value within six to twelve months. Others see customers moving ahead of other regions on RISE with SAP. Joel sees two distinct cohorts behind those views.

The first cohort moved their SAP estates to the cloud early and has since invested in innovation around the core. When SAP pitches RISE with SAP and access to Joule, their response is often to ask why, because they are already happy with the strategy they have built. These customers may already be running SAP S/4HANA in cloud environments but are not necessarily using Cloud ERP or have a subscription-based contract.

The second cohort sees genuine value in RISE with SAP but approaches it as experienced cloud adopters. “They have a business case and know what they want to achieve,” Stehr said. “And as part of this, they’re going to tick off the projects that they are most interested in along the way.”

From Pipes and Plumbing to Data and AI

The way AWS shows up in customer conversations has changed dramatically over the past two years, and especially in the last six to twelve months. Historically, customers saw AWS as the provider of “pipes and plumbing,” which Stehr acknowledged is an undifferentiated proposition if that is all a customer looks at. The value conversation has now moved to what customers can do with their SAP data once it lands on AWS, and that is broadly about data and AI.

To support this, AWS has invested heavily in a suite of modernization agents for SAP customers, which fall into two categories. The first helps customers migrate from legacy SAP platforms, including SAP ECC, SAP BW and SAP PI/PO, all of which are approaching the end of mainstream maintenance in 2027.

The second is a toolkit to help SAP customers build agentic applications. Stehr noted that customers still on SAP ECC have so far not had any access to Joule, leaving them asking what ERP looks like for them in an agentic world. Meanwhile, some RISE with SAP customers simply don’t want to be tied to one vendor’s agentic platform. “What they want is something agnostic that spans their entire enterprise application environment,” he said, “so they can orchestrate across and between their different applications.”

This matters because SAP’s own roadmap is still landing. The vision announced at SAP Sapphire is compelling, but the newer Joule capabilities like AI Assistants forming part of the Autonomous Suite remain limited to early adopters, and customers on SAP ECC have yet to see what will be available to them. Stehr was clear that SAP will eventually make this available for most customers. The question is when, and he does not expect it within the next three to six months.

The Cost of Waiting

For Stehr, the most important conversation customers should be having internally is not about which technology to adopt. It is about the cost of not starting now. “If someone sits and waits for SAP to deliver, and let’s use three months as an example, what is their competitor able to deliver in market in those three months?” he asked. “If you start three months late, you’re not going to be three months behind, because AI is not linear. It’s exponential.”

The initiatives AWS sees customers pursuing tend to fall into three categories:

• Incremental improvement to internal processes. These organizations have likely  recognized that agentic capabilities would complement, not change, their SAP processes, and decided to start moving forward.
• Internal process transformation. An example of this type of customer is one switching on RISE with SAP and exploring how agentic AI can reduce its reliance on external application management services, keep ticket volumes down after go-live, and monitor integrations between SAP and Salesforce.
• External experience transformation. These represent larger, more ambitious initiatives aimed at reshaping customer or citizen experiences.

Stehr is also fielding a question he says he answers more often than he can count: why not simply build our own system and tools? As agentic coding accelerates, the build-versus-buy debate is more open than it has been in years. Stehr describes the ANZ SAP base as largely “buyers rather than builders,” yet some customers are now choosing to build rather than renew or buy new applications. His advice is to weigh carefully the risk of replacing SAP against retaining it for core processes and investing in transformation around the edges.

What This Means for SAPinsiders

Quantify the cost of waiting. Before deferring AI initiatives until SAP’s roadmap lands, estimate what competitors could deliver in the short term. Progress in agentic AI for SAP compounds, so building internal maturity now through small, value-focused projects may matter more than waiting for the perfect platform.

Put cloud investments to work on data. If an SAP estate already runs in the cloud, look beyond infrastructure to the data and AI services available around it. SAP ECC customers in particular may want to explore vendor-agnostic options for agentic applications that span SAP and non-SAP systems, while making a sober build-versus-buy assessment for core processes.

Connect with AWS in Sydney. Mastering SAP Collaborate runs 18–19 November at ICC Sydney. Meet Joel Stehr and the AWS team to discuss modernization agents for legacy SAP solutions, AI solutions like Amazon Bedrock, and how agentic toolkits can start or accelerate building today.