Joule Agents

Enterprise

SAP

Product overview

Name of Agent: Joule Agents
Short description of agent: "Joule Agents are AI agents embedded into every business function and accessed with role-based assistants. They use SAP’s process expertise to automate complex workflows and deliver AI value at scale." (link, archived)
Date of release: 09/10/2024 (link, archived)
Advertised use: "use built-in business process context and integration with your data and applications to automate complex workflows" (link, archived)
Monetisation/Usage price: included in SAP cloud subscriptions. (link, archived) Base price not available.
Who is using it?: enterprises
Category: Enterprise

Company & accountability

Developer: SAP
Name of legal entity: SAP SE
Place of legal incorporation: Walldorf, Germany (link, archived) (German legal registration number: HRB 719915)
For profit company?: Yes
Parent company?: SAP
Governance documents analysis: Terms of Use (link, archived)
AI safety/trust framework: Responsible AI at SAP (link, archived)
Compliance with existing standards: SAP has achieved certification for ISO/IEC 42001, (link, archived)

Technical capabilities & system architecture

Model specifications: This (link, archived) says "Select from the latest AI models to equip agents with powerful planning and reasoning capabilities."
Observation space: Business data stored in Joule Studio, and tool call returns.
Action space: Access to many tools, details (here, archived).
Memory architecture: According to this, "Retains intermediate results and past interactions, ensuring continuity and statefulness across multi‑step workflows."
User interface and interaction design: Wizard-like UI for adding context, connecting data sources, assembling teams of agents, etc.)
User roles: Designer (user is designing an agent to deploy within SAP's Joules)
Component accessibility: Closed source

Autonomy & control

Autonomy level and planning depth: L1: user has full control over how to design the agent. L1-3 for resulting agent designed by the system, offers many opportunities for the user to steer its trajectory and decide what to do with outputs
User approval requirements for different decision types: Resulting agent seems to regularly request user feedback and involvement
Execution monitoring, traces, and transparency: Agent doesn't seem to show CoT, but does regularly presents the user with options for how to proceed
Emergency stop and shut down mechanisms and user control: User can decline to approve agent actions
Usage monitoring and statistics and patterns: Unclear, but assume it's available through SAP Joules

Ecosystem interaction

Identify to humans?: None found
Identifies technically?: - "Joule Studio is a capability in SAP Build that allows organizations to create Joule agents" (link, archived) - SAP Build Apps have designated IP addresses (link, archived)
Interoperability standards and integrations: - MCP support, A2A ((link, archived), (link, archived))
Web conduct: None found

Safety, evaluation & impact

Technical guardrails and safety measures: SAP states the Joule Studio agent runtime includes “built-in guardrails for data protection” (runtime-level; not described as a builder-configured option) and human-in-the loop. Both just high level descriptions without details (link, archived)
Sandboxing and containment approaches: None found
What types of risks were evaluated?: None found
(Internal) safety evaluations and results: None found
Third-party testing, audits, and red-teaming: None found
Benchmark performance and demonstrated capabilities: None found
Bug bounty programmes and vulnerability disclosure: Yes (link, archived)
Any known incidents?: None found
← Back to 2025 Index