AI Agents: Join fan-out, overlapping totals and a custom measure service in SAP CAP

Using a small AI agent eval to explore double-counting and measures.

How Companies Are Approaching MCP: Reflections From Amsterdam

This blog explores how companies are putting MCP(Model Context Protocol) into practice, drawing on insights from AGNTCon + MCPCon Europe. It covers protocol evolution, tool discovery, skills and agent execution, alongside our work on Open Resource Di...

AI Product Management: Not Every Decision Needs a Language Model

Three weeks ago a logistic regression beat eight LLMs at reading job postings. I gave it a rematch against a model built for tables. TabPFN-3.5 beat the lion, beat GPT-6 Astra and Claude Fable on the same postings, ran in 12 seconds per thousand, cos...

WebMCP in a Fiori app: what happens when the job gets hard?

Agent-ready Fiori: does WebMCP pay off for an AI Agent?

System-1 decision models and tabular foundation models: who decides what in a business process

Enterprise workflows are weighed down by thousands of routine micro-decisions—routing a support ticket, verifying a supplier edit, or flagging an invoice for review. Rigid rule engines become brittle as edge cases multiply, while general-purpose LLMs...

Joule Learning Journal: October 2026

Here are 5 learning resources that helped me get started with SAP Build and Joule Studio this month. 

Why I stopped letting LLMs decide what to fetch

Letting the LLM decide what to fetch works. It's also non-deterministic, sequentially slow, and hard to audit.A short note on declarative pre-fetch as the default for enterprise agents and where tool calling is still the right answer

Building CAP Agents for Knowledge Work - Announcing Two Vector-Store CDS Plugins

A CAP Agent can continue a conversation and, with a long-term Memory Store which I made 2 weeks ago, can remember useful information across conversations. But a support agent also needs to answer questions from product guides, troubleshooting notes, ...

GeoTrack: tracking trips in SAP HANA Cloud

A personal location pipeline on SAP HANA Cloud spatial, CAP, RAP and Fiori Elements. Built with Claude Code.

My productivity is boosted with Joule work desktop, this is the reason why

There is a long-standing debate about the use of IA and productivity in the corporate environment. This article explores the real-life experiences of an AI Consultant and how the mix of experience, collaboration, and reskilling generated outstanding ...

AI had one set of instructions and broke it fearlessly — then apologised!

Hi. I'm Andréa Molina, an SAP engineer with 20+ years of experience, working hands-on with BTP, UX, and AI. I shared my  An Engineering Manifesto for AI-Assisted SAP Development — Let's Build It Together blog with 19 principles, examples and code pat...

An AI agent that rebooks passengers in ABAP, and waits for your approval first

Trying the CAP-level agents from the September 2026 CAP release with a custom RAP service.

An Engineering Manifesto for AI-Assisted SAP Development — Let's Build It Together

AI helps us learn and write code. We have been shipping code faster than ever. Someone with no IT background can create software now. But software is not just code. It is structure, security, governance and much more. We deal with sensitive data. An ...

Teaching an AI agent to draw SAP BTP diagrams the SAP way

My AI agent could draw boxes, but not SAP. So I gave it the SAP BTP icons and the SAP diagram rules. Meet ARC-DRAW, an open-source MCP server, and see how you can help make it better.

Migrating from PI-Sheets to Execution Sheets: What You Need to Know Before 2030

With the announcement that PI-Sheets fall under SAP’s compatibility scope—and therefore usage rights expire at the end of 2030—customers are now faced with an important question: What is the right path forward for existing PI Sheet implementations?

When Agents Multiply: Human Oversight in the Autonomous Enterprise

When we started this blog post series, we asked a hard question: how do we keep humans meaningfully involved as AI systems get more autonomous every day? Today, that question is more urgent than ever. SAP’s vision of the Autonomous Enterprise promise...

Running Custom ABAP AI Scenarios? Don't Let Certificate Renewal Catch You Off Guard!

Are you building or even running a business ABAP Cloud AI scenario as described in Swathy's blog about the ABAP AI SDK powered by ISLM?Embedding Generative AI capabilities directly into ABAP Cloud applications has never been more accessible, and the ...

Take your AI skills further with the learning zone for SAP TechEd 2026

 With SAP TechEd 2026 coming up, explore learning opportunities to build your skills in SAP Business AI and Joule. Discover the future of work with Joule and deepen your understanding of AI reasoning, foundation models, and enterprise knowledge. Cont...

Building CAP Agents That Remember - Announcing `cap-agents-memory` Plugin

With the September 2026 release of CAP, CAP now has a first-class plugin for building AI agents directly within CAP applications. You can annotate a CDS service with @AGent, expose it through the A2A protocol, use CAP service capabilities such as act...

Stop Getting Bad AI Outputs: A Prompt Engineering Guide for SAP Consultants

A Beginner's Guide — For Every SAP Consultant, Every Module, Every Role

Governance at Scale: Why a Clean Core Is What Makes Agent Governance Enforceable

SAP has spent 2026 building governance infrastructure for AI agents — a verification workflow through AI Agent Hub, policy enforcement through MCP Gateway. Both promise trust badges and audit trails. Neither can deliver on that promise against an obj...

Smart Report - Example with VA05N (List of Sales Orders)

This proof of concept (POC) shows how to turn standard SAP reports into interactive AI tools. While tested using the VA05N sales report, this exact setup can be used for any other SAP report. It connects SAP data to a local AI, allowing users to chat...

Who Am I Acting As? Identity Propagation from SAP BTP to Multi-Agent Systems

A conversational walk-through for SAP BTP developers — from Cloud Connector principal propagation to agent-to-agent (A2A) delegation.Think back to the last BTP app you built that talked to an on-premise S/4HANA system. Chances are, every backend call...

ABAP AI's Three Circles in Practice

Sonja Liénard, SAP's head of ABAP platform, describes ABAP AI as three concentric circles — ABAP Cloud at the center, developer tooling around it, agents on the outside. It's a clean diagram. This part tests it against the messy case from Part 4: thr...

From AI Assistance to Human-on-the-Loop: A Compliance Detection Case Study

What if your compliance team could stop chasing false alarms and start focusing on what really matters? That’s exactly what happens when AI moves from being a helpful assistant to a tool that automatically handles routine cases.

Version Control and Backup of SAP Cloud Integration Artifacts with Git - Step-by-Step Guide

When working on SAP Cloud Integration, it is a good practice to keep a backup of your integration flows and other artifacts outside the tenant. Git helps us do this in a simple way. Every time we make a change, we can save a version in a Git reposito...

Deploying a Docker Model to SAP Generative AI Hub

Organizations often need to deploy proprietary AI and machine learning models that are not available as standard offerings. In this blog, you'll learn how to deploy a Dockerized model to SAP Generative AI Hub, covering image creation, registry setup,...

Side-by-Side Without Relocating the Debt: Governance Patterns for CAPM (Series 3, Part 4)

Part 3 showed how SAP's Custom Code Migration Agent handles the mechanical half of clean core work. This part addresses what the agent explicitly hands back to a human: the objects that get side-by-side extended into CAPM. Done under deadline pressur...

Inside the S/4HANA Custom Code Migration Agent: A Technical Walkthrough (Series 3, Part 3)

Part 2 walked through assessing clean core manually with ATC and Project Kernseife. This part goes inside SAP's own answer to the same problem: the S/4HANA Custom Code Migration Agent, generally available since Q2 2026. It's not a chatbot bolted onto...

How did I extend my Joule Work Desktop (JWD) and how do Guinea Pigs relate to that…

Trust Isn't Enough: Validating Human-in-the-Loop for Oversight

Trust Isn't Enough: Validating Human-in-the-Loop for OversightWhen we deploy an AI model, we measure everything. From accuracy, precision, recall, confidence scores to benchmark performance. Having these numbers gives us the confidence that our syste...

Grounding Joule Agent on Your Own Documents Without Object Store

A hands-on walkthrough: bring your own AWS S3 bucket, ground it through SAP AI Core's Grounding Management, and expose it to a custom Joule agent with an SAP Build Action.The full flow at a glance — from your S3 bucket to a grounded Joule agent.   If...

Autonomous Enterprise Partner Learning Week | October 12–16, 2026​

SAP Partners: Get ready for SAP TechEd with five days of expert-led learning, hands-on technical sessions, and certification preparation focused on the Autonomous Enterprise and the latest SAP innovations.

Where the Seconds and the Dollars Go in a Text-to-SQL Agent for SAP Data

We instrumented every phase of a text-to-SQL agent running on an SAP schema. The step we assumed was expensive turned out to be the cheaper one, and the cost was accumulating somewhere we were not looking.

From Business Accelerator Hub to a Conversational Agent: Building Joule Studio Skills Step by Step

Build a conversational "Order Health Assistant" using only standard SAP APIs — no custom development. This step-by-step walkthrough covers finding the right APIs on Business Accelerator Hub, setting up a destination, wrapping APIs as Actions, and bui...

Assessing Clean Core: A Hands-On ATC + Kernseife Walkthrough (Series 3, Part 2: Clean Core + AI)

RISE with SAP customers get the RISE Methodology Dashboard in Cloud ALM as a standardized starting point for clean core assessment — and that's genuinely the right first step. This part goes one layer deeper: the ATC mechanics actually powering that ...

SAP-RPT-1.6, Tabular Orchestration, and RPT Playground API: Now Available

SAP-RPT-1.6, Tabular Orchestration and a higher RPT Playground API limit are now live. Together, these cover the full arc from experimentation to production. 

More than a Checkpoint: Human-in-the-loop in Practice

How can we define mandatory control points in the AI agent’s decision-making chain?What does it actually mean for a human to be "in the loop"? In our first blog post of this series, we introduced the concept of human oversight in AI systems and explo...

A SAP Datasphere task chain monitoring agent

1. BackgroundNothing was on fire. The task chains ran, mostly successfully, and when they didn't someone eventually noticed.The problem was a question. Our BI team lead has downstream data integration work that depends on a handful of chains completi...

August Developer Challenge - NOW CLOSED

The August Developer Challenge for the new Joule Studio is now closed.

Abapsmith - let your LLM work with ABAP code

Abapsmith is a new open-source MCP server that lets an LLM read, write, activate, run, test, and debug ABAP on a live system over ADT.

Why Clean Core Became an AI Precondition (Series 3, Part 1: Clean Core + AI)

Clean Core used to be sold as ERP upgrade hygiene. In 2026, it became something closer to a precondition — the RISE with SAP deals closing this year (Airbus, Shell, Morgan Stanley, Eli Lilly) aren't just modernizing infrastructure, they're building t...

Managing SAP AI Core using PowerShell

A lightweight PowerShell Sample for managing SAP AI Core resource groups, orchestration configurations and deployments.

Why Human Oversight Matters More Than Ever and What It Means

With the advent of AI agents, there is increasing interest in a topic that had been around for a long time in the AI Ethics world: Human oversight and determination. But why did it become so important now, and what does it mean?

SAP EMEA Business AI Agent Hackathon 2026 - Our Runners-up Medal Winning SAP Joule Agent

Background and Context – Why we need an agent or Why we choose this use case for the AI HackathonAs we transition legacy Microsoft AX entities into SAP-managed S/4HANA Private cloud operations, invoice traceability has become increasingly complex. A ...

WebMCP: Giving AI Agents a Structured Interface to the Web

A deep-dive into WebMCP, Chrome's proposed standard for exposing structured tools to AI agents directly from web pages, covering the imperative and declarative APIs, React integration, real-world patterns, and what it means for the future of agent-na...

AI-Powered Forecasting with SAC, SAP RPT & Gen AI - Part 1: Predictions That Explain Themselves

The Problem This POC SolvesPredictive forecasting in financial planning isn't a new concept. SAC has offered predictive scenarios for a while, and traditional ML has always been possible. So why hasn't every planner adopted it?Because the barrier to ...

August Developer Challenge Week 4 - Vercel

This is Week 4 of the August Developer Challenge on Joule Studio with a focus on Vercel integration.SAP and Vercel have partnered to enable developers to build lightning-fast, modern frontends (using React and Next.js) for enterprise applications and...

From Predictions to Real Decisions: How SAP RPT-1.5 and Joule Agents Orchestrate AI Workflows

A deep-dive into building AI workflows in Joule that turn tabular structured data into key supply chain decision engines using SAP-RPT-1.5 Tabular AI model. 

August Developer Challenge: A look at n8n with Vibe Coding

n8n workflows, as you saw in the Week 3 challenge, can be defined by JSON text file, and therefore you can easily vibe code to create complex and useful workflows. I will show you what I did do manage the challenge itself.