SAP expands Joule into an agentic work layer as Autonomous Enterprise goes live
SiliconANGLE Kyt Dotson
SAP is turning Joule into a full-screen work layer for apps and agents. It’s now live, and SAP wants humans to steer the bots, not vanish from the loop.
Based on reporting by SiliconANGLE, Kyt Dotson — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
SAP is pushing Joule past chatbot duty and into the middle of enterprise work. At SAP Connect 2026, the company said its Autonomous Enterprise architecture, first introduced at Sapphire in May, is generally available this month alongside Joule Work and Desktop.
That matters because SAP is not selling autonomy as a switch you flip. Manoj Swaminathan, SAP’s president and chief product officer of Autonomous Suite, said the idea is a journey: people need to observe agents, describe what they want, and understand how the software will use their data before handing over real work.
The technical pitch is simple enough to say and messy enough to matter. SAP wants “AI on apps” rather than “AI in apps.” Joule Work is meant to sit across SAP and non-SAP software, while Joule Desktop follows the customer across screens and pulls together enterprise context from different systems. Swaminathan used total corporate spend as the example: ERP, travel and expense, and procurement data can be combined into one answer instead of forcing a person to stitch it together.
Joule itself is getting a bigger role. SAP said the assistant, first launched in 2023, is now a full-screen experience rather than a plain chat client. With the Autonomous Enterprise framework going generally available, it will also pick up updated agents for finance, supply chain, spend, workforce and customer experience, plus other industry-specific functions.
SAP is also making a broader argument about why its version should work. Swaminathan said enterprise customers already bring the data and cloud lakehouses; SAP organizes that data into a structured knowledge graph, with ERP semantics, policies and relationships intact. He contrasted that with general-purpose models and agents from OpenAI Group PBC and Anthropic PBC, which do not supply that enterprise structure on their own. The company says the software will be consumed through AI units, with customers able to see how those units are used.
My take — AI-written commentary, not fact-checked reporting
This is the part of enterprise AI that actually makes sense: keep the human in charge, wrap the model in governance, and stop pretending a chat box is a strategy. SAP’s pitch is basically that the boring plumbing matters more than the demo, which is refreshingly unsexy and probably correct. The rest of the market can keep worshipping copilots until the bills arrive.
Read more about this at: SiliconANGLE
