SAP Joule Studio 2.0 — rolling out June 2026 — is not just an upgrade to an agent builder. It is a fundamentally new development model: describe a business outcome in plain language and the platform generates a full artifact suite — PRD, technical specs, code scaffolding, evaluation harness, and multi-agent orchestration — automatically. Here is the complete technical and strategic guide for enterprise teams deciding how to build on it.
What Is SAP Joule Studio 2.0 — And Why Is It Different From Everything Before It?
SAP has been building AI tooling for developers since the early BTP days. What makes Joule Studio 2.0 — announced at SAP Sapphire in May 2026 and beginning its rollout in June 2026 — categorically different from every previous iteration is a single architectural decision: intent-based development.
In every previous SAP AI development environment, you started with a blank canvas and built. You chose a model, designed a prompt, wired up tools, created an agent definition, wrote evaluation logic, and configured the deployment runtime. Each of these steps required expertise — either in BTP configuration, SAP AI Core, LLM prompt engineering, or agent orchestration frameworks. The result was a platform that worked well for experienced AI developers but created a steep barrier for the business analysts, process owners, and SAP functional consultants who understand the business problem most deeply.
Joule Studio 2.0 inverts this. You describe what you want the agent to do — in plain language — and the platform generates the complete implementation stack: the product requirements document, technical specifications, code scaffolding, workflow logic, evaluation suites, and multi-agent orchestration definitions. SAP CTO Philipp Herzig demonstrated this live at Sapphire 2026: a single natural-language prompt producing a deployable multi-agent system in minutes.
This is not a feature update. It is a new development model.
Intent-Based Development: How It Actually Works
The intent-based development flow in Joule Studio 2.0 operates in four stages:
Stage 1: Describe the Outcome
The developer — or business analyst, or process owner — describes the desired business outcome in natural language. Not a technical specification. Not a prompt template. A business description: "I need an agent that monitors our open purchase orders in SAP S/4HANA, identifies any orders where the supplier confirmation is more than 3 days overdue, sends a Joule notification to the relevant buyer with the overdue order details, and escalates to the procurement manager if the delay exceeds 7 days."
That is the entire input. Everything else is generated.
Stage 2: Generated Artifact Suite
Joule Studio 2.0 processes the natural-language description and generates a complete artifact suite:
- Product Requirements Document (PRD): A structured specification of what the agent should do, what data it needs, what actions it should take, and what success looks like — available for review and editing before any code is written
- Technical Specification: The implementation design — which SAP APIs the agent will call, which Joule Skills it will invoke, how the orchestration logic flows between steps, and what error handling is required
- Code Scaffolding: The actual agent code — in the developer's target framework (SAP CAP, n8n, LangChain, AutoGen, LlamaIndex) — populated with the specific API calls, skill invocations, and business logic derived from the PRD
- Evaluation Suite: Test cases, expected outputs, and edge case scenarios generated automatically from the PRD — so the agent has a complete quality gate before deployment
- Multi-Agent Orchestration Definition: If the described outcome requires multiple specialised agents working together, the orchestration layer — handoff protocols, data contracts between agents, fallback paths — is generated alongside the individual agent definitions
Stage 3: Review, Refine, and Extend
Every generated artifact is editable. The intent-based generation is a starting point, not a black box. Developers review the PRD for accuracy, adjust the technical specification to match enterprise conventions, modify the generated code for performance or style requirements, and extend the evaluation suite with additional test cases. The platform preserves the connection between the PRD and the implementation — so changes to the PRD propagate to the technical specification and vice versa.
This review-and-refine model is what distinguishes intent-based development from the fragile "generate-and-hope" pattern of early AI code generation. The generated artifacts are not a one-time output — they are a maintained, living specification that evolves with the agent.
Stage 4: Deploy to Managed Runtime
Deploying the finished agent requires zero infrastructure decisions. The Joule Studio managed runtime — detailed below — handles all execution, scaling, monitoring, and security automatically. The developer clicks deploy; the agent is available within Joule's enterprise engagement layer.
The Four Solution Types You Can Build in Joule Studio 2.0
Joule Studio 2.0 supports four distinct solution types, each addressing different enterprise automation needs:
1. Agents
Autonomous AI agents that execute multi-step business tasks: gathering data from SAP systems, reasoning over it, taking actions, and escalating exceptions. These are the core Joule agents — the same architecture that powers SAP's own 30+ pre-built Joule agents for finance, procurement, HR, and supply chain. Custom agents built by enterprises or partners in Joule Studio run on the same runtime and are invocable through the same Joule interface as SAP's native agents.
2. n8n Workflows
Process automation workflows built using the n8n open-source workflow automation framework, now natively integrated into Joule Studio. n8n's visual, node-based workflow builder is already familiar to thousands of enterprise automation teams — Joule Studio's n8n integration brings SAP ERP actions into n8n workflows as first-class nodes, without requiring custom API integrations. For enterprises that already run n8n for non-SAP automation, this means a single workflow platform can now orchestrate processes that span SAP and non-SAP systems.
3. Agent Extensions
Extensions to SAP's existing pre-built Joule agents, adding enterprise-specific business logic without replacing the underlying agent. If SAP's standard Procurement Agent handles purchase order creation but your enterprise has custom approval rules, tiered supplier classifications, or industry-specific compliance checks, an Agent Extension adds those specific behaviours to the standard agent without rebuilding it from scratch. Extensions are upgrade-safe: when SAP releases a new version of the base agent, extensions adapt without breaking.
4. SAP CAP Applications
Full custom business applications built using SAP's Cloud Application Programming (CAP) model — the same framework SAP uses internally for S/4HANA extensions and BTP applications. CAP applications are the most flexible solution type, enabling enterprise development teams to build complex, stateful business applications with custom data models, rich UIs, and deep ERP integration. For complex requirements that exceed what an agent or workflow can address, CAP applications in Joule Studio provide the full development power of BTP with Joule Studio's AI-assisted scaffolding accelerating the build.
The Managed Runtime: Zero Infrastructure, Enterprise-Grade Security
The managed Joule Studio runtime — available through Early Adopter Care now, GA targeting Q3 2026 — eliminates the infrastructure burden that has historically been a barrier for mid-market enterprises and citizen developer programmes. Here is exactly what "managed" means in technical terms:
SAP AI Core Foundation
All agents run on SAP AI Core — SAP's enterprise AI orchestration engine, the same infrastructure that powers Joule across all 35 SAP solutions where it is currently deployed. AI Core handles model access (GPT-4o, Claude, Gemini, and SAP's own foundation models), request routing, load balancing, and model version management — none of which the developer needs to configure or maintain.
NVIDIA OpenShell Isolation
Each agent runs in an isolated, sandboxed environment provided by NVIDIA OpenShell, with configurable guardrails that prevent agents from accessing data or taking actions outside their defined scope. This is critical for enterprise governance: an agent built by a business analyst for procurement automation cannot inadvertently access HR data or finance records. The sandbox boundaries are defined in the agent's permission scope at build time, not enforced only at runtime — making the governance model auditable and reviewable before deployment.
SAP HANA Cloud Memory
Agents have access to persistent, long-term memory powered by SAP HANA Cloud. This is the technical foundation that enables agents to maintain context across multi-session interactions: a procurement agent can remember that a specific supplier was flagged for delivery delays last quarter, and apply that context to a new order decision today. HANA Cloud memory is governed, structured, and query-optimised — not a raw vector store — which means agent memory is searchable, auditable, and consistent with SAP's enterprise data governance standards.
Built-In Observability
Every Joule Studio agent automatically receives metering, distributed tracing, security monitoring, logging, audit logging, and telemetry from the managed runtime — without any developer configuration. For enterprise compliance teams, this means every agent action is logged and auditable from day one of deployment. For operations teams, this means agent performance and error rates are visible in SAP Cloud ALM dashboards automatically.
The No-Code to Pro-Code Spectrum
One of the explicit design goals of Joule Studio 2.0 — and a direct response to feedback that Joule Studio 1.0 was too constrained for professional developers — is supporting the full development spectrum without mandating a specific approach:
No-Code (Citizen Developers and Business Analysts)
Business users describe their automation intent in natural language. The platform generates a complete agent. The user reviews the PRD for accuracy and deploys. No code is written, reviewed, or maintained. This is the "citizen developer mode" — enabling the people closest to the business process to build agents without IT dependency. The governance guardrails (NVIDIA OpenShell scope, SAP AI Core permissions, HANA Cloud access controls) ensure that citizen-built agents are safe for enterprise deployment even without developer review of the generated code.
Low-Code (Functional Consultants and Power Users)
SAP functional consultants and power users — who understand SAP process logic deeply but are not professional developers — use Joule Studio's visual workflow canvas and the n8n integration to build and modify agents with minimal coding. The generated code scaffolding is a starting point they refine through the visual interface rather than direct code editing. This is the development mode for the majority of SAP implementation partners building standard enterprise use cases.
Pro-Code (Professional Developers and Architects)
Professional developers work directly in Visual Studio Code or Cursor using the Joule Studio Code Editor extension (SAPSE.joule-studio-code-editor, available on the VS Code Marketplace). The extension provides AI-guided scaffolding, contextual code generation, and intelligent recommendations for SAP APIs, Joule Skills, and CAP model patterns. Developers can work with any supported agent framework — LangChain, AutoGen, LlamaIndex — without being locked into SAP's native orchestration model. Generated artifacts from the intent-based engine are editable at the code level with full IDE tooling support.
Joule as the Enterprise Engagement Layer: The Strategic Shift
Joule Studio 2.0 sits within a larger strategic repositioning that SAP formalised at Sapphire 2026. Joule has been recast from a natural-language chat interface into SAP's enterprise engagement layer — the primary interface through which employees, managers, and external systems interact with SAP's AI capabilities under what SAP calls the Autonomous Enterprise strategy.
As of Q1 2026, Joule is:
- Live across 35 SAP solutions — from S/4HANA Finance and EWM to SuccessFactors, Ariba, and Concur
- Powered by 30+ specialised agents covering finance, procurement, supply chain, HR, and IT service management
- Backed by 2,500+ Joule Skills — discrete, reusable capabilities that agents combine to execute complex business workflows
The custom agents enterprises and partners build in Joule Studio 2.0 run alongside — and interoperate with — all 30+ SAP native agents through the same engagement layer. A custom procurement compliance agent built by a financial services firm using Joule Studio can hand off to SAP's native Ariba Sourcing Agent and receive results back, with the entire interaction visible in a single Joule interface. This is the ecosystem architecture that the €100 million partner fund is designed to populate with industry-specific and function-specific agents.
What Enterprises Should Do Now: A Practical Readiness Roadmap
Joule Studio 2.0 is rolling out to Early Adopter Care in June 2026, with GA targeting Q3 2026. For enterprise teams planning their approach, the following sequence maximises readiness:
- Inventory your highest-value automation candidates: Which repetitive, rules-based processes consume significant time from skilled employees? Which exception management workflows have clear escalation logic that could be automated? Document these before touching the platform — the intent-based development engine works best when you arrive with a clear business outcome description.
- Assess your Joule Skills coverage: Review which of the 2,500+ Joule Skills are relevant to your process candidates. Agents built on existing skills deploy faster and require less custom code maintenance. Where there are gaps, assess whether Agent Extensions to existing SAP agents cover them, before committing to building a net-new agent.
- Identify your citizen developer candidates: Which functional consultants, business analysts, or process owners understand your target processes most deeply? These are the natural candidates for citizen developer agent building — not the central IT team. Building a small cohort of 3–5 citizen developers who can build and maintain department-specific agents is more sustainable than centralising all agent development in IT.
- Join Early Adopter Care for the managed runtime: The 12-month free design-time access for EAC participants is the most cost-effective way to build organisational experience with Joule Studio before the GA commercial model kicks in. SAVIC can assist with EAC enrolment and initial agent design.
- Plan your governance model before your first deployment: The NVIDIA OpenShell permission scopes, AI Core access controls, and HANA Cloud data access boundaries need to be defined as a governance policy before the first citizen-built agent goes live. Retrofitting governance after deployment is significantly harder than defining it upfront.
SAVIC's Joule Studio 2.0 Practice
SAVIC's SAP Business AI practice has been working with Joule Studio since its initial GA and is actively evaluating the 2.0 capabilities through the Early Adopter Care programme. Our Joule Studio practice covers:
- Use Case Identification and Prioritisation: Structured workshops to identify the highest-ROI agent automation candidates in your SAP landscape, assessed against Joule Skills coverage and data readiness
- Agent Design and Build: End-to-end agent development using the intent-based engine — from natural-language specification through PRD review, code scaffolding, evaluation suite configuration, and managed runtime deployment
- Citizen Developer Enablement: Training and coaching programmes for functional consultants and business analysts to build and maintain their own department-level agents, with governance guardrails designed for enterprise deployment
- Governance Framework Design: Definition of agent permission scopes, HANA Cloud data access policies, AI Core model access controls, and Cloud ALM monitoring dashboards — establishing enterprise-grade governance from day one
Contact SAVIC for a Joule Studio 2.0 readiness assessment — we help enterprise teams move from the Sapphire announcements to production agents with a clear business case, technical architecture, and governance framework.
Frequently Asked Questions
How does SAVIC approach SAP implementation projects?
SAVIC follows a structured One Piece Flow methodology — delivering SAP projects in focused, iterative waves that reduce risk, accelerate time-to-value, and keep business disruption minimal. Each phase is scoped, tested, and signed off before the next begins.
What industries does SAVIC serve with SAP solutions?
SAVIC serves 12+ industries including manufacturing, automotive, consumer products, retail, life sciences, chemicals, oil & gas, real estate, and financial services — across India, UAE, Singapore, the US, UK, Nigeria, and Kenya.
How long does a typical SAP S/4HANA implementation take with SAVIC?
Timelines vary by scope. GROW with SAP public cloud deployments can go live in 8–12 weeks using SAVIC's pre-configured accelerators. Full RISE with SAP private cloud transformations typically take 6–18 months depending on landscape complexity, data migration volume, and custom code remediation.
Does SAVIC provide post-go-live SAP support?
Yes. SAVIC's MAXCare managed services programme provides post-go-live application management, Basis & infrastructure support, continuous improvement, and defined SLA-backed support across all SAP modules — with 24/7 coverage options for critical production environments.


