Artificial Intelligence is transforming how we design and extend SAP applications. Many SAP professionals are now exploring how to start working with AI inside the SAP ecosystem.
In this post, I share my hands-on experience with SAP AI Core, from my first steps to building a real Proof of Concept (POC) using low-code tools like Build Apps and BPA.
Since May, I’ve been working on a POC with SAP AI Core. At that time, I knew nothing about AI. Step by step, by practicing with the examples and samples provided by SAP, I managed to build several POCs integrating SAP AI Core with low-code tools.
The main objective is simple: accelerate development cycles and innovate with custom applications. On my roadmap, I focused on using low-code solutions like SAP Build Apps and Process Automation (BPA), while integrating AI agents to speed up processes.
Maybe in a future post I will share more details about the POCs I realized. But today, my first goal is to encourage SAP technical consultants to start this journey.
As a Senior SAP Technical Consultant, I realized that integrating an AI agent for a first POC or when consuming partner models does not require brand-new skills. But if the objective is to build and train your own model, then of course new competences are needed.
With SAP AI Core, we can easily integrate partner models (like Azure OpenAI, ChatGPT, etc.) through APIs. For most SAP technical profiles, it’s enough to follow a simple end-to-end example:
Select a partner model (text, image, etc.) in SAP AI Launchpad
Deploy it and get the URL
Configure the destination and start consuming your AI agent
Another option in SAP AI Launchpad is the Prompt Editor, which lets you create your own prompts on top of partner models. The key point here is security: your data is not stored in the model — it’s managed securely by SAP.
This means that integrating external AI agents is feasible without heavy upskilling.
But in my opinion, the real power of SAP AI Core comes when you train your own models. Of course, this requires upskilling in new environments and languages. The key here is HANA Vector combined with RAG (Retrieval-Augmented Generation). Together, they enable customers to enrich their knowledge databases and build custom agents that are more efficient, optimized, and deliver a brilliant user experience.
I believe that SAP AI Core will become a central part of the future SAP landscape, enabling faster innovation and smarter business processes.
I’d love to hear your thoughts and experiences with SAP AI Core. Have you already built a POC or started exploring it?
Bluesky
