SAP Joule: From Launch to 2,400 Skills — How SAP's AI Copilot Evolved

SAP Joule went from zero to 2,400+ skills across 13+ applications in under three years. It now supports 40+ autonomous agents, adoption grew ninefold through 2025, and 67% of SAP’s Q4 cloud orders included Business AI (Futurum Group, 2026).

Nobody saw that trajectory coming. Most people (us included, if I’m honest) wrote it off as a neat conversational add-on when SAP first showed it at TechEd in 2023.

TL;DR: SAP Joule evolved from a single-purpose Q&A tool in 2023 to a multi-agent platform with 2,400+ skills, 40+ agents, and Joule Studio for custom agent building (GA Q1 2026). Adoption grew 9x in 2025, with real-world results including 3x faster development (GigaOm, 2025) and 50% faster sales data retrieval.

What Was Joule at Launch?

When SAP announced Joule at TechEd in September 2023, enterprise AI adoption sat at around 55% globally (McKinsey, 2025, measuring from the 2023 baseline). Joule entered as a natural-language interface for asking questions about your SAP data. Type a question, get an answer from your business systems. That was it.

And that first version was pretty limited. It worked mainly within SAP SuccessFactors for HR queries. Narrow scope, pre-trained responses. No agents, no workflow automation, no cross-application anything. A smarter search bar, if we’re being honest.

But the architecture was interesting. SAP built Joule on BTP with a modular design that could connect to any SAP cloud application. And they went multi-LLM from the start (dynamically selecting from OpenAI, Anthropic, Google models), which meant Joule wasn’t tied to one AI provider’s limitations. That was a proper bet. Not obvious at the time, but it paid off.

How Did Joule Evolve Through 2024?

Deloitte’s 2026 State of AI survey of 3,235 leaders found that 66% of organisations report productivity gains from AI, but only 34% use it for deeper transformation (Deloitte, 2026). SAP’s 2024 roadmap for Joule was clearly aimed at that deeper tier. We watched three major shifts happen, quarter by quarter.

Expansion to core business applications

Joule moved beyond HR into S/4HANA, Sales Cloud V2, Service Cloud V2, and Commerce Cloud. Each integration brought domain-specific skills. Not just answering questions anymore: creating purchase orders, generating opportunity summaries, drafting follow-up emails. A different animal entirely.

Contextual business awareness

Early Joule could answer “What were last quarter’s sales?” The 2024 version could answer “Which accounts are at risk of churning based on declining engagement?” That shift from retrieval to analysis changed the value proposition. Suddenly it was doing work, not just fetching data.

Prebuilt skill library growth

By mid-2024, SAP reported over 1,300 prebuilt skills. Each one a specific business action Joule can execute: generating job descriptions in SuccessFactors, summarising customer interactions in Sales Cloud, drafting procurement documents. The catalogue kept growing.

What Changed in 2025 That Made Joule Matter?

2025 was the inflection point. SAP Business AI reached 34,000 customers, about 60% of cloud customers actively using AI features and another 20% in implementation (SAP News Center, 2025). I reckon three things drove that acceleration.

From copilot to agent platform

The biggest architectural shift was Joule Agents: autonomous AI agents that handle multi-step workflows without needing a human nudge at each stage. By Q4 2025, SAP had deployed 40+ agents across finance, procurement, HR, and CX.

The measured impact is solid:

• Cash Management Agent: up to 70% reduction in manual effort
• Utilities Self-Service Agent: up to 90% reduction in customer contact costs
• Performance Preparation Agent: 50% less prep time, 80% fewer follow-ups
• Sales and Service Agents: up to 50% productivity improvement

These aren’t experimental features buried in a lab. They’re documented in SAP’s Q4 2025 release notes (SAP News Center, 2026).

Ninefold adoption growth

Joule customer adoption grew 9x through 2025. Of SAP’s 50 largest Q4 deals, 90% included AI or SAP Business Data Cloud. Two-thirds of Q4 cloud orders included Business AI, up 20% from the previous quarter (Futurum Group, 2026).

What drove the jump? Partly the agent capabilities. But I think it was also SAP’s decision to include Joule in cloud subscriptions with a free usage tier rather than selling it separately. Smart move. Remove the procurement barrier and adoption follows.

Joule Studio: build your own agents

Joule Studio reached general availability in Q1 2026. Low-code environment for building custom agents. This is the shift from “use what SAP gives you” to “build what your business needs.”

A procurement team can now create an agent that reads incoming RFP documents, extracts key requirements, cross-references supplier capabilities, and drafts response templates. All through Joule Studio, no custom ABAP required. That’s a de facto democratisation of enterprise AI. And it happened faster than most of us expected.

What Concrete Results Are Companies Getting?

A survey of 1,600 executives across eight countries found a 16% average return on AI investments, expected to nearly double within two years (SAP/Oxford Economics, 2025). The IBM Institute for Business Value dug deeper: SAP gen AI ROI currently averages 6.8%, projected to reach 12.2%, with mature users seeing 20% average profit margins versus 16% for peers (IBM IBV, 2024).

The developer story stands out. A GigaOm benchmark study showed SAP Build with Joule achieved 3x faster development speed compared to traditional methods, with a 30% boost in development task completion (GigaOm/SAP, 2025). Java and JavaScript developers using Joule in SAP Build Code cut development time by 30%. That’s not a slide deck number. That’s people shipping code faster.

In SAP Sales Cloud V2, Joule helps sales teams navigate data up to 50% faster through natural-language queries, opportunity summaries, and lead scoring. We covered the hands-on experience in our Joule guide for Sales Cloud.

How Does Joule Compare to Other Enterprise AI Tools?

McKinsey’s 2025 survey found that 88% of enterprises use AI, but only about 6% qualify as “AI high performers” (McKinsey, 2025). One reason: most AI tools live outside the business processes where decisions actually happen. That’s the gap Joule targets.

It comes down to integration depth.

General-purpose AI copilots (ChatGPT, Claude, Gemini) are spot on for content generation and open-ended reasoning. But they operate outside your business systems. They don’t know your purchase order history, your pipeline, or your inventory levels unless you feed that data in manually.

Productivity copilots (Microsoft Copilot, Google Duet) integrate with office suites: documents, spreadsheets, email. Helpful for knowledge work. Disconnected from ERP, CRM, and supply chain processes.

Joule operates inside the business context. It knows your data model, respects your permissions, executes within the process flow. The tradeoff: it won’t write your blog post or debug your Python code. It does one thing, and it does it well: make SAP processes faster and smarter.

SAP has also built bi-directional integration between Joule and Microsoft 365 Copilot. Create Teams channels or schedule Outlook meetings directly from SAP applications through Joule. Neat bridge between productivity tools and business processes.

What Should You Consider Before Investing in Joule?

Despite 88% AI adoption, Deloitte found that 37% of organisations remain at “surface level” with AI, minimal process changes (Deloitte, 2026). Moving past that surface level with Joule takes honest prep work.

Data quality comes first

Joule’s insights are only as good as the data underneath. If your CRM has duplicate accounts, inconsistent opportunity stages, or sparse activity logging, Joule will surface unreliable recommendations. Clean your data before you turn on AI features. We see this trip people up constantly. It’s the number one thing that goes wrong.

Start with one high-confidence feature

Don’t try to deploy 40 agents at once. Pick one feature with clear ROI: natural-language queries in Sales Cloud V2, automatic case classification in Service Cloud, or cash application in S/4HANA. Run it for 30-60 days, measure the impact, then expand. Crawl, walk, run.

Licensing works differently than you’d expect

Joule comes included in SAP cloud subscriptions with a free usage tier. Beyond that threshold, you purchase SAP AI Units based on annual message consumption. Budget planning should account for usage growth as adoption spreads. This catches people off guard regularly.

Process clarity matters more than the technology

Nota bene: before enabling Joule Agents, map the workflows you want to automate. Every step. Every decision point. Every exception. Agents can’t improve processes you don’t understand yourself. If you can’t whiteboard it, an agent can’t run it.

FAQ

How many skills does SAP Joule have?

As of Q4 2025, Joule includes 2,400+ prebuilt skills across 13+ SAP applications (SAP News Center, 2026). Up from roughly 1,300 in mid-2024. Each skill is a specific business action: generating reports, creating purchase orders, drafting customer follow-up emails.

What are Joule Agents and how many are available?

Autonomous AI agents that handle multi-step business workflows without needing a human prompt at each stage. Over 40 agents are available across finance (Cash Management Agent), HR (Performance Preparation Agent), service (Utilities Self-Service Agent), and sales. They come as part of SAP Business AI.

Is SAP Joule free to use?

Included in SAP cloud application subscriptions with a free usage tier. Beyond the included threshold, overages require purchasing SAP AI Units based on annual message consumption. The 350+ embedded AI features come with cloud subscriptions.

What’s the difference between Joule and Joule Studio?

Joule is the AI copilot with 2,400+ prebuilt skills for existing processes. Joule Studio (GA Q1 2026) is a low-code agent builder for creating custom agents tailored to your specific workflows. No ABAP or custom development needed.

What ROI can companies expect from Joule?

Varies by use case. Documented results: 3x faster development with SAP Build (GigaOm, 2025), 70% reduction in manual cash management effort, 50% faster sales data retrieval, 90% reduction in customer contact costs. SAP-specific gen AI ROI averages 6.8%, projected to reach 12.2% (IBM IBV, 2024).