With the launch of Autonomous Enterprise, SAP is expanding Joule into an agent work layer

SAP SE today announced a shift from the promised artificial intelligence autonomy of agents to defining a path for users to interact with agents to work outside of apps, not just within them.

During SAP Connect 2026, the company said it is making the Autonomous Enterprise architecture, introduced at Sapphire in May, generally available this month along with Joule Work and Desktop. with an expanded suite of autonomous assistants and agents.

“When we talk about autonomous enterprise, we realize that it’s going to be a journey for the customer, and it’s not something that you can say will be autonomous tomorrow just because you have these agents and assistants now,” said Manoj Swaminathan, SAP president and chief product officer of the Autonomous Suite, in an interview with SiliconANGLE.

Joule, SAP’s generative AI assistant, launched in 2023 and has become the company’s autonomous workhorse. Today the company announced numerous upgrades that make it a powerful automation, productivity and development tool that acts as a conversational interface layer for customers, working alongside apps rather than within them.

According to Swaminathan, “autonomous” doesn’t mean just handing everything over to the agents and then letting them go; It means giving people the tools to observe, describe, and trust agents before delegating processes to them. This starts with the ability to control, direct and understand how agents work with their data, apps and corporate context before releasing agents, and the transparency to see and monitor their activities.

“The first principle is that it is AI in apps, not AI in apps,” Swaminathan explained.

The focus is Joule Work, which provides the framework to integrate AI assistants into customers’ workflows, enabling autonomous work with both SAP and non-SAP applications. A user looking at a screen has already seen AI agents working in apps where they can open a word processor and have an agent write them a memo or an email program and have an agent look up previous conversations. But when an agent understands enterprise resource planning and connects financial documents and email conversations without having to switch between the two, it becomes SAP’s domain.

For example, Swaminathan gave an example of overall corporate spending. Instead of treating the question as belonging to an application, Joule might determine that the answer requires information from ERP, travel and expense systems, and the procurement platforms. These data sets could then be combined and the result presented as a business response.

This interface system is called Joule Desktop and is located where the customer uses their apps and accesses the same company information. It provides an intelligent layer that can perform tasks related to the work itself instead of tying it to resources internal to the application.

“Joule is no longer a natural language chat client for us; it’s a full-screen experience,” Swaminathan said.

With the general availability of the framework, Jule will receive updated agents for finance, supply chain, expenses, HR/customer experience, and other industry specializations across the autonomous suite. This gives customers access to comprehensive capabilities and agents with expertise, knowledge and business acumen who can process corporate data.

Transitioning from AI models to autonomous companies

According to Swaminathan, enterprise customers bring their data and cloud lakehouses with them; SAP organizes this data and provides it with a structured knowledge graph. General-purpose models and agents from OpenAI Group PBC and Anthropic PBC cannot provide this when simply inserted over enterprise data. ERP data, processes, semantics, policies and relationships are all important – this is the “secret sauce” that SAP brings to the market.

SAP is not alone in this trend. Salesforce Inc. and ServiceNow Inc. are also driving the industry trend toward headless applications, context layers and agent control layers. Enterprise AI is becoming less about adding co-pilots to the applications themselves and more about building an intelligent layer on top of the customer experience.

Swaminathan emphasized that the opportunity for SAP customers lies in the possibility of a letter of intent; The agent decides which systems to include and then does the work. Humans stay informed, maintain control, auditability and transparency, and agents are treated like employees in the system, including security, systems of record and execution. Agents are another layer above the old application interface, acting as intelligent communication tools on behalf of people.

“Autonomous entrepreneurship is no longer a vision,” Swaminathan said. “It’s real, with general availability in October.”

SAP will continue to use AI units as the primary commercial monetization for these capabilities. The model will be primarily consumption-based rather than purely subscription-based, with agent inference units consuming and SAP providing the customer with visibility into that usage.

The Joule updates and other autonomous enterprise features will be available later in October and customers can already see agents in SAP’s Discovery Center.

Image: SiliconANGLE/Microsoft Designer