Log in to save this article and keep your favorite resources in one place.
For more SAP utilities insights, sign up to receive ASUG’s free The Topic: Utilities newsletter, and to get involved and keep up with the latest, join ASUG’s Utilities Community. Connect with SAP teams across the utilities community as they explore AI adoption, clean core, data strategy, and more at SAP for Utilities, Presented by ASUG in San Antonio, October 6 – 9, 2026.
As SAP’s SVP and General Manager for Energy & Natural Resources Industries, Torsten Welte is responsible for translating SAP’s strategic direction into actionable steps for the energy sector, which includes utilities. A month after SAP Sapphire 2026, Welte spoke to ASUG from Germany, fresh from a run of discussions with utility CIOs at a system integrator leadership conference.
Those conversations kept circling the same gap: between what SAP announced — the autonomous enterprise, agentic AI, a tool chain now built around Joule — and what a utility actually has to do to realize it. On the cost side, SAP has moved aggressively. At Sapphire 2026, SAP introduced agent-led migration tooling it says can cut ERP migration effort by more than 35% — automating system analysis, code remediation, configuration, and testing.
Welte is optimistic about the technology, but just as quick to say what has to come first. In the interview, Welte explained the leadership role of SAP as a technology partner for the utilities industry, where SAP’s AI investments deliver value for utilities, what CIOs should prioritize in the near-term and long-term, and the extent to which organizations can trust autonomous agents within critical infrastructure.
The following interview has been edited and condensed for length and clarity.
Q: How should utility CIOs quantify and prioritize the business value of SAP’s recent Sapphire announcements — cost optimization, reliability, customer experience?
There were many announcements at Sapphire, especially around AI. The autonomous enterprise was announced, along with partnerships such as those with Palantir and the first Palantir service partner, Accenture. Then the data-side acquisition of Reltio was announced, along with industry AI and many AI-based tools improving IT operations. Customers can see that SAP is entering the agentic world and becoming their agentic partner with a focus on value.
What does that mean? From an operational perspective, if I’m a CIO, what’s in it for me? Looking at the entire tool chain, the supported transformational and operational tools all focus on cost reduction and faster value delivery. In Business Transformation Management (BTM) and other areas, we are infusing AI throughout the transformation — from requirements gathering to execution, configuration, value management, testing, consulting with Joule, and development with the new Joule Studio. Joule Studio makes development easier for everyone, especially on the business side, so you don’t need expensive technical developers anymore. On the cost side, we’ve done a lot. Christian announced a significant 30% reduction in implementation costs with BTM and other tools. This is helping all utility CIOs reduce cost and increase the speed to valuable innovations.
Another critical piece is the agentic side. We announced Industry AI and are partnering with selected customers to expand our portfolio in this space. We have worked and will work on agents and other capabilities to improve critical end-to-end processes. One of the first ones relevant for utilities and other asset-intensive companies aims for autonomous asset management. This is an example of the agentic space where we are doing much more than before.
Q: S/4HANA, AI, and BTP are the top focus areas for utilities. What pain points is SAP hearing in each, and how is it responding?
On the BTP side, the feedback was on how to accelerate development and get people, especially on the business side, to use BTP as a platform. Activities around Joule Studio v2, which combines Build and the Joule Studio into one tool, make consuming BTP services easier. They are also working on integration, and the official MCP servers were released for customers to interact with the SAP environment.
On the AI side, the biggest initial feedback was how to manage capabilities, what is available, and how the commercial model works. ECC customers asked if they can use Joule and what AI can do for them.
Christian’s announcement allowing ECC customers who are still on the journey to RISE — not finished with it — to use Joule and other AI capabilities is critical. They can begin an AI strategy on a signed RISE contract without completing the transformation first. BTP now offers many AI capabilities through specific services. Also, many early features are now part of the base capability and do not consume AI units. We are expanding higher-value AI capabilities.
Every meeting, CIOs asked, “What AI tools can you help me with to really get into the transformation quicker, faster, and with less cost?” The announcements now put significant AI-infused tools into the hands of IT and the business.
Q: The utilities roadmap spans meter-to-cash, customer service, and asset management. What’s the single most important action a utility should take in the next 12 months?
My response echoed what many CIOs at this week’s conference stressed. To accelerate processes and deliver more business value, you need a solid foundation based on good data. If data is siloed or inconsistent, AI will hallucinate and behave unpredictably. To quote Peter Hinssen, “There is more fiction written in Excel than in Word.”
To put an identity layer across data and the enterprise data model, you need good master or transactional data in a system that accesses all data pockets. BDC will be critical. It may not be strictly necessary, but it greatly improves efficiency by adding security and federation. Plus, it accelerates AI.
Another key piece is a solid foundation with S/4 and RISE. I believe in RISE because, as a CIO, having one partner providing processes, operations, and infrastructure as a service with an SLA lets me operate differently with the business, especially with AI. I have common processes and transactions, but execution can vary through the agentic layer, enabling variances and hyper-customization. This is not possible in the normal ECC world due to technical restrictions like system locking. Plus, many of the ECC systems have extensive modifications that are not known to AI.
Q: How is SAP addressing asset-intensive challenges — aging infrastructure and the workforce shortage, the so-called Silver Tsunami?
I love that because I’m part of the silver generation. The response starts with announcements around Distributed Energy Resources (DER). This is transforming the utilities industry and requires much change management and new technologies. There aren’t many established players in this space yet, so we’re bringing the right players together and helping the community understand what this technology change means.
But that change is in lockstep with organizational changes, so understanding the customer now, not only from a bill perspective, but all the way from a consumption pattern to a supply pattern, is going to be super critical. Providing tariffs to the community based on demand and supply is going to change completely.
End customers today only focus on — what was my meter reading, what was my bill, and the bill disputes. It now goes into a whole new set of details on your EV charger, solar panel, and so on, which the customer wants to manage through the agent on the phone, the virtual agent, and the website alike. Every channel you interact through has to work off the same data and processes, and you have to coordinate it.
You have to interact with the silver tsunami community all the way down to the Gen Zs that expect to just talk to the phone and get all that information right there, not even talking to somebody physically.
For the aging workforce, we need to capture their knowledge, much of which is already in system configurations. By applying AI and new innovations on top, we can harvest documented and undocumented capabilities. We also provide training through community work, bringing content and information to constituencies like enterprise architects and user communities. There are many tools to interact differently, including a new learning platform.
As a community, we probably need to do more. The SAP AI tools offered now should also help attract young talent. The offering of learning platforms, experience centers, and learning as you go can definitely help address the silver tsunami.
Q: For utilities still navigating S/4HANA adoption, how is SAP evolving the roadmap to reduce complexity, cost, and risk — and to deliver business outcomes, not just technical modernization?
The AI-infused transformation tool set is a major area. We have the business transformation tool chain, including Signavio and LeanIX, so companies can measure current operations. Anytime processes change, you can measure the impact. The same applies when deploying agents or AI — you can use LeanIX to see the impact, how processes improve.
The second focus is on industry AI and value-driven domain discussions. We are not doing AI just for technology’s sake but to create business value. On the asset side, unplanned downtimes have a big impact on power generation or on power distribution. Optimizing maintenance resources is critical. Additional focus is also on reducing inventory and reducing safety issues. We overlay the agentic layer on the data and processes. This allows us, for example, to re-optimize existing processes based on the new parameters to adapt to the new situation.
You can see we heavily invest in products, but we are also changing how we engage customers, with forward-deployed engineering to create and deploy the agentic layer. SAP no longer works in the back office with a six to twelve-month wait for new features. We develop capabilities with customers and roll them out faster, shortening time to value because these have been developed and proven by customers.
Q: In terms of total cost of ownership, what structural changes is SAP making to reduce the long-term cost curve across infrastructure, licensing, and the implementation of S/4 and AI-enabled architectures for utilities?
My hypothesis is that running your own servers is going to be very expensive, even if some are already paid off, because of the upkeep and the cyber risk. The hardening alone is significant, and it spans everything from a standard utility to one with nuclear capabilities, where you need completely different infrastructure, requirements, and certifications. We are investing heavily in those areas where some of our competitors are not, and making sure we adhere to our SLAs and improve them over time. That results in a lower cost for customers.
We have also talked about the tools and the transformation effort around the tool chain. There is a set of clean-core tools coming out that will help significantly from a deployment and transformation perspective.
And lastly, our industry AI area will move toward a different commercialization model, one that is much easier and centered on what value means rather than pure cost.
Q: What guardrails or reference architectures is SAP putting in place so customers can make durable investment decisions rather than continuously pivot?
If I were a customer, I would focus on RISE in this area. I don’t have to wait through 12 to 18-month governance cycles for new releases, then add another six to twelve months to provide capabilities to the business. Cybersecurity in RISE is much stronger than in your own environment. We have invested in data, partnerships, and embedded AI capabilities, which should excite CIOs and the business.
Just look at the Gen AI Hub: how many different LLMs are in there, and how often the versions change. If you had to manage the releases with your own environment, certifying them, testing them, benchmarking them, and onboarding them, that’s quite a bit of work.
The other piece is the capability of Joule, centered around the knowledge graph. That knowledge graph has the entire data model of SAP in it, and has the context of the data model in it, so they can operate in that environment, which you will understand. That getting extended is the key factor, how businesses can accelerate their own development, the onboarding of AI capabilities, and scale this to a different level.
Q: As AI agents begin executing end-to-end processes, how does SAP ensure utilities can trust an agent to write to production, and what governance prevents unauthorized autonomous actions on critical infrastructure?
We’ve discussed the new API policy, which some people got upset about, but if you really look at it, it hits home exactly in that area. With MCP servers, you can access systems, if they’re not guarded and governed in the right way, in a completely different way: drag down performance, trigger unauthorized transactions, and perform write-backs.
The policy we put in place is significant. We believe that without working closely with customers here, volatility will increase, costing them much more. Having to catch and defend against cybersecurity issues means system downtime will have a greater impact. Overall, cybersecurity is the top priority at SAP.
And lastly, critical infrastructure, especially regulated ones with nuclear operations, we offer capabilities with SAP NS2 for customers who look for a stronger, harder, more cybersecurity-sensitive organization.
Q: If you were a utility CIO today with an existing SAP estate, what two or three moves would you prioritize over the next 24 months?
My view matches what I heard from 25 CIOs at a major system integrator leadership conference. They all said, first, have a clean core and infrastructure that allows you to scale.
Second, ensure data is in order — not just clean, but the right amount — and provide access across different systems of record.
Third, which ties to innovation through AI but is even more critical, is change management. Prepare the organization for changes from an agentic perspective and AI adoption. Don’t underestimate that some groups fear losing jobs to AI rather than being enriched by it. Adopting AI instead of resisting it is crucial, along with a solid value foundation driving AI and innovation.
For more SAP utilities insights, sign up to receive ASUG’s free The Topic: Utilities newsletter, and to get involved and keep up with the latest, join ASUG’s Utilities Community. Connect with SAP teams across the utilities community as they explore AI adoption, clean core, data strategy, and more at SAP for Utilities, Presented by ASUG in San Antonio, October 6 – 9, 2026.
You Might Be Interested In
Log in to save this article and keep your favorite resources in one place.
Log in to save this article and keep your favorite resources in one place.
Log in to save this article and keep your favorite resources in one place.
Log in to save this article and keep your favorite resources in one place.