“SAP TechEd 2025: The Agentic AI Era Begins – Key Highlights and What They Mean for Developers”
November 7, 2025
SAP | Technical Write-Ups

SAP TechEd 2025: The Shift to Agentic AI

SAP TechEd 2025 marks a pivotal evolution — from integrating AI to architecting with Agents. With Joule as the central intelligence, SAP’s ecosystem is becoming more open, modular, and agent-driven.

Agentic Evolution in SAP

  • Joule Agents → Context-aware, extensible, enterprise-ready AI Agents. 20+ Agents are already available, and by the end of this year, the count will be 40+..
  • Agent Builder in Joule Studio → Build or extend SAP-delivered Agents.
  • Pro-Code Agents → Develop in any programming language (AWS, Microsoft, LlamaIndex, etc.).
  • Agent2Agent (A2A) Protocol → Open standard for interoperability; partners include AWS, Microsoft, Google, ServiceNow.
  • Agent Orchestration for working with multiple agents.
  • Agen Governance using SAP LeanIX AI Agent Hub
  • Agent Tracking & Benchmarking using Agent Mining in Signavio
  • Agentic Memory in HANA Cloud → Persistence and recall for Agents interacting with enterprise data.

This year, SAP formalizes the Agentic paradigm — moving from LLM-powered assistants to enterprise-grade autonomous agents integrated into SAP’s data and process landscape

SAP AI Foundation: Models, Tools & Intelligence

  • New foundation models (GPT-5 Pro, Gemini Live 2.5, Qwen 3, Claude 4.5 Haiku, Magistral Medium).
  • Perplexity is integrated into the AI Foundation for enhanced retrieval and reasoning.
  • Prompt Optimizer for cross-model tuning (in collaboration with Not Diamond)
  • Evals Framework to measure LLM performance.
  • SAP-RPT-1 – The first SAP foundation model for structured/tabular data. Outperforms traditional AutoML for structured data tasks.
    • “RPT” = Relational Pre-trained Transformer.
    • Open source version: SAP-RPT-OSS (available on Hugging Face).

SAP-RPT-1 is a game-changer for enterprises heavily dependent on structured data — bridging the gap between LLM reasoning and relational intelligence.

Data Fabric and Interoperability

  • SAP Business Data Cloud connectors for Databricks, Snowflake, and Google Cloud.
  • SAP Snowflake → Snowflake AI & data capabilities as a native SAP solution extension.
  • Data Product Studio → Transform raw data into governed, reusable data products.

The data story is about openness and collaboration — SAP isn’t locking down ecosystems but rather connecting them.

Developer Empowerment: Open Tools, AI SDKs, and MCP

  • SAP Cloud SDK AI → Build AI-native applications.
  • MCP (Model Context Protocol) Servers → for CAP, ABAP, and UI5.
  • VS Code Extensions for SAP Build & ABAP Cloud.
  • AI-driven ABAP (Cloud) → with upcoming SAP-ABAP-1 LLM.
  • n8n Integration → Low-code automation meets enterprise AI.
  • Joule for Developers → Integrated into SAP Build, CAP, Fiori, and Mobile.

SAP doubles down on developer experience — meeting coders where they are, whether in SAP Build, VS Code, or external agentic tools like Claude, Cursor, or Windsurf.

Integration, Automation, and Quantum

Key highlights:

  • Integration Suite MCP Gateway Support → Create and expose MCP servers for third-party systems.
  • AI-driven anomaly healing and auto-migration of integrations.
  • Joule-generated iFlows → Intelligent automation in integration design.

SAP is gradually layering intelligence into integration and even exploring the quantum frontier.

Conclusion – The New SAP Stack

Quantum Computing initiative → Exploring business problem-solving via quantum algorithms and making the Qunatum capabilities accessible to customers.

SAP’s roadmap weaves together Agentic AI, structured-data intelligence, and developer freedom — redefining what it means to build, extend, and automate in the enterprise world.

Subscribe to Our Blog