Michael Ameling, President of SAP BTP and Member of the Extended Board, SAP
As organisations rethink their technology landscapes in the age of AI, SAP is sharpening its focus on BTP as the engine for enterprise innovation. Michael Ameling, President of SAP BTP and Member of the Extended Board, shares insights on the platform's key pillars, the shift toward AI-native architecture, and what's driving strong AI adoption in India.
Can you walk us through the BTP segment and key pillars of SAP's AI-native strategy?
My current focus is on the extension portfolio, specifically SAP Build, which allows us to build custom applications. Even our own SAP applications, like the next-generation business networks, supply chain orchestrator and intake management, are developed with that. These solutions are built on BTP and level the full AI native stack.
The second large pillar is the SAP Integration Suite. Ninety per cent of the developers deal with integration challenges. The question is how to bring process data together. We have our iPaaS solution, an enterprise-wide integration solution with many open connectors.
We have embedded AI everywhere in the SAP product portfolio within the SAP Business Suite, like in job description generation and optimising processes in finance. Delivering our agents and assistants is exciting; we're harnessing GenAI and large language models (LLMS) to enable autonomous workflows that progressively improve finance, ERP, human capital management and more.
Customers have heterogeneous data landscapes. They have data lakes, SAP systems, critical data and sometimes copy data back and forth -- same data, multiple silos, not harmonised. At the beginning of the year, we launched SAP Business Data Cloud, which solves this problem.
We can now integrate third-party solutions and have an offering that brings this data together without copying it back and forth, because copying is expensive, leads to delays and data loss.
We decided not to develop our own model because it's a high investment. It was better to partner with the best vendors since we don't know the best for each use case. Some models are better for finance, some for HCM and some for cost efficiency. We partnered with a company called NotDiamond -- a prompt optimiser -- to develop one prompt and change the model depending on performance, cost or sovereign/legal restrictions.
From the user experience perspective, we wanted conversational interaction with systems. This is where we started with Joule, our conversational co-pilot. Meanwhile, Joule developed into an orchestrator of Business AI. If you have a request like "What is my revenue prediction for next quarter?", it goes to the finance system. Same for agents. Joule orchestrates which agent to use for which use case. Over time, we will bring more agents.
To sum it up, we offer a full AI-native stack that is future-ready. Customers can invest today and keep innovating without starting from scratch. Even if you're on a legacy system, you can start on BTP, develop use cases, evolve, integrate with other data sources through Business Data Cloud and Integration Suite and have agent-to-agent communication.
What is the nature of your customers adopting BTP -- is it mostly enterprises?
It varies. Large enterprises use BTP in all dimensions -- Integration Suite as enterprise integration, hundreds of apps built on top of BTP by large IT departments.
Mid-sized customers are also doing quick innovative use cases because BTP runs in multiple data centres globally. It comes with all enterprise requirements. Once it's approved by IT and security, the next use case is easier. It has data privacy, protection and compliance and is attractive for mid-size companies.
Partners also develop entire whitespace solutions on top of BTP and can cover the entire market because they have tight integration into SAP systems.
What is the percentage of the overall customers using BTP?
We see over 30,000 customers on BTP adopting this in all dimensions. Every customer who will be in the cloud and every RISE customer, for example, gets SAP Build and therefore, BTP embedded. In the future, every cloud customer may have access to BTP because it's natural -- extensibility and integration come out of the box. It's not for all legacy customers because they're still in the transformation, but they are heading there.
What role does India play in SAP's global BTP and AI strategy?
India has the right talent and topics here and continuous growth to deliver the entire stack. Not only does AI -- India contribute across all dimensions and is a major pillar of SAP delivery.
We also have a strong partner ecosystem and a booming customer base. We have good relations with hyperscalers locally. Multinationals come here and we show use cases.
Our recent survey showed that in India, 93 per cent of businesses see a return on AI investment over the next 3 years, far above the global average. Globally, companies spend roughly $27 million on AI; in India, it's $31 million, showing commitment.
We delivered 20 agents during Connect, with more coming. We have teams developing end-to-end and a forward-deployed engineering approach, working with customers to develop use cases together.
There's been an ongoing debate about the lack of clear ROI from AI investments. From your discussions with customers, what measurable outcomes are emerging?
We did an analysis on value generation from AI. It's different company to company and depends on the stage they are at. You can introduce AI, but not gain value because your data might not be right, or because you're still in transformation. If processes are not optimised, it's hard to gain value. There are customers in transformation, whom we help with our RISE methodology to fit the standard.
Cloud frees capacity for innovation -- customers don't need to maintain systems and can focus IT effort on innovation and AI use cases. Return on investment comes over time. Early investors see ROI sooner -- customers like JP, ABB and others already see returns. Those who wait may miss learning opportunities and won't know how to apply AI or its prerequisites.
Comments
Published on November 28, 2025
READ MORE