SampleSampletypeA worked example you read or clone. Evidence: it demonstrates, it never provides.See every one →What this type is →by SAPby SAPoriginPublished by SAP itself. A visible facet, never an input to a reading — a dormant official plugin reads exactly like a dormant independent one.See every one →Reference

teched2025-AI165

An SAP TechEd 2025 hands-on that builds a custom agent in Joule Studio's Agent Builder and has it collaborate with an external partner agent over the Agent2Agent (A2A) protocol.

A fictional Business Partner Invoicing scenario runs through Joule Studio's Agent Builder: either attach one of its out-of-the-box tools to an agent or build your own with Joule Skills, which is also what agent collaboration uses. The Business Partner Invoicing Agent answers from the records of known business partners, creates tickets in ServiceNow to have new ones added, and attaches an invoice to the business partner record in SAP S/4HANA, which the README marks simulated. Where it cannot find the business partner in question, it collaborates in the background with IBM's ServiceNow Ticketing Agent, invoked through IBM watsonx Orchestrate, over the Agent2Agent (A2A) protocol. Joule, SAP's AI Copilot, is what the goal list calls the primary entry point; the agent itself is fictional, built on a pre-release Agent Builder against systems and a machine the session provides.

Facts

Last pushLicenceApache-2.0

Where it sits

Lives in
Demonstrates
Agent toolingBTPAgent toolingcapabilityBuilding an agent into your own application — the tools it may call, the state it keeps between turns, and the framework it runs in. Distinct from MCP servers, which expose an application to somebody else's agent, and from agent guidance, which is knowledge handed to a coding agent.
Requires
Joule StudioJoule Studioartifactwhere the custom agent and its Joule Skills are built; the session grants access to it and to its Agent Builder.Node.jsNode.jsoutside the atlasone of the few things the exercises need on the machine, at the version the README names; the session's machine has it installed already.

Who is behind it

SAP SamplesSAP SamplesOrganisationMaintains this
In the atlas: maintains 222
@SAP-samples
maintainerAjay SorengPersonContributes to this
In the atlas: contributes to 36
@ajaysoreng
julian-schambeckPersonContributes to this
In the atlas: contributes to 6
@julian-schambeck
Kay SchmitteckertPersonContributes to this
In the atlas: contributes to 6
@kay-schmitteckert
Werner TodPersonContributes to this
In the atlas: contributes to 1
@wernertod
4 contributors

What it declares

Every artifact

Type
Sample
Origin
By SAP
Use as
A reference

As a sample

Form
Workshop or CodeJam material

Connections3

README

quoted from SAP-samples/teched2025-AI165 under Apache-2.0, read
SAP-samples/teched2025-AI165/README.mdOn GitHub

AI165 - Interconnect custom AI agents with your cloud and partner ecosystem

REUSE status

Important

Welcome to the Agent Builder Lab Preview!

You are working with a pre-release version of the Agent Builder in Joule Studio. This gives you an early look at our upcoming capabilities. Please keep the following in mind:

  • Features Are Subject to Change: The user interface (UI), terminology, and functionalities you see in this lab may differ from the final, generally available (GA) product.
  • For Educational Use Only: This environment is designed for learning and experimentation, not for production use.
  • Potential Instability: As a preview version, you may encounter occasional instability or minor bugs. The exercises are designed to work with the current state of the platform. If you get stuck, please notify a session instructor.

This repository contains the material for the SAP TechEd 2025 session called AI165 - Interconnect custom AI agents with your cloud and partner ecosystem.

Goal formulation for this hands-on

By the end of this session, participants will have successfully:

  • Built a custom agent in Joule Studio using its new Agent Builder
  • Enabled the agent to collaborate with an external partner agent from IBM
  • Utilized the Agent2Agent (A2A) protocol for standardized communication between agents and enhanced agent interoperability
  • Invoked the custom agent from within Joule (SAP's AI Copilot) as the primary entry point

The scenario: Business Partner Invoicing

Important

The Business Partner Invoicing Agent described here doesn’t exist and is purely a fictional concept created for a hands-on session scenario.

In this hands-on session, we will work with a fictional Business Partner Invoicing scenario in which your company's business partners send invoices for services they have provided. Assume the invoices have been sent via email, for example. The invoice then needs to be attached to the corresponding business partner record in SAP S/4HANA so that it can be paid later. The Business Partner Invoicing Agent assists you with this task and is able to:

  1. Provide information about business partners from the records of known business partners
  2. Create tickets in ServiceNow to add new business partners to the records
  3. Attach an invoice to the business partner record in SAP S/4HANA (simulated)

To build custom agents like the one described above, you use Joule Studio, a capability within SAP Build, and its new built-in Agent Builder. You can either attach one of the out-of-the-box tools to an agent or build your own using Joule Skills. The latter can also be used for agent collaboration, where an agent collaborates with another (external) agent. Joule, SAP's AI Copilot, can then leverage your custom agents and their skills to complete designated tasks.

In this scenario, you as the end user interact only with the Business Partner Invoicing Agent, which maintains knowledge about the business partners. However, if the agent cannot find the business partner in question, it collaborates with IBM's ServiceNow Ticketing Agent (invoked via IBM watsonx Orchestrate in the cloud) in the background to open a ServiceNow ticket requesting business partner creation.

IBM watsonx Orchestrate is IBM’s open platform for building, managing, observing, and governing AI agents. In addition to its low-code and pro-code agent development capabilities, it also provides a catalog of prebuilt agents that simplify access to AI tools from various domains like Human Resources, Sales, Procurement, and Customer Care. This workshop leverages the latter, integrating ServiceNow customer care capabilities like ticket management. Reference: https://www.ibm.com/docs/en/watsonx/watson-orchestrate/base

Note that the two agents communicate using the Agent2Agent (A2A) protocol to facilitate agent interoperability. This workflow minimizes manual work and accelerates the storage and processing of invoices that your company must pay.

Architecture Diagram

Solution Diagram Architecture is subject to change

Requirements

There are only a few requirements to follow the exercises. Everything will already be installed on the machine provided for the live hands-on session. You will also be given access to Joule Studio and its Agent Builder.

  • A Chromium-based browser (e.g., Google Chrome)
  • Node.js v22.17.0 (LTS)
  • A text editor, such as VS Code
  • Access to Joule Studio and its Agent Builder

Exercises

Contributing

Please read the CONTRIBUTING.md to understand the contribution guidelines.

Code of Conduct

Please read the SAP Open Source Code of Conduct.

How to obtain support

Support for the content in this repository is available during the actual time of the online session for which this content has been designed. Otherwise, you may request support via the Issues tab.

License

Copyright (c) 2025 SAP SE or an SAP affiliate company. All rights reserved. This project is licensed under the Apache Software License, version 2.0 except as noted otherwise in the LICENSE file.

Apache-2.0, read

Details

Sources3
Added by handdiscovered herechecked
github.com/SAP-samples/teched2025-AI165tree, pushes, contributors, CIchecked
README.mdquoted, and what the entry declareschecked

Sources are graph entities: the same page may back several entries, and an unreachable source flags every row that rests on it.

Compatibility and provenance
discovered viaAdded by hand
first seen in the atlas
ecosystem and domainbtp, ai-and-agents
Contributors4

From the repository’s contributor list, .

History in the atlas1
  1. First seen.
For your agent
Ref
artifact:sap-samples-teched2025-ai165
JSON
/artifacts/sap-samples-teched2025-ai165.json
MCP
Connecting an agent