SAP and Google Cloud agents
- SAP and Google Cloud expanded a partnership to deploy multi‑agent AI with Gemini Enterprise as the orchestration hub. - The arrangement emphasises zero‑copy data access so agents can act across systems without heavy duplication. - Early marketing use cases include deployable 'Joule Agents' for campaign optimisation, with first customer availability expected in H2 2026, signalling a focus on workflow economics. ( )
SAP and Google Cloud said on April 22 they are expanding their partnership so companies can run multiple artificial intelligence agents across both vendors’ software, with Google’s Gemini Enterprise coordinating the work. (sap.com) The setup links SAP Engagement Cloud, SAP Customer Experience and Joule with Gemini Enterprise, which SAP and Google said will serve as a central hub for data integration and agent coordination. SAP said the first example is marketing, where agents can execute campaign plans from high-level goals set by a user. (sap.com) Both companies are pitching “zero-copy” access to data, meaning the systems can use information where it already sits instead of moving large datasets into a new store first. SAP and Google said SAP Business Data Cloud Connect for Google BigQuery will provide bidirectional zero-copy access with security and governance controls. (sap.com, cloud.google.com) The announcement lands as Google rolls out Gemini Enterprise as an agent platform for businesses, with tools to build, run and govern agents in one system. Google introduced that platform at Google Cloud Next ’26 on April 22. (cloud.google.com, cloud.google.com) SAP has been building the other side of that stack around Joule, its generative artificial intelligence assistant, and Joule Agents, which SAP describes as ready-to-use agents for automating business workflows. The companies have been widening that work since at least April 2025, when SAP said it was advancing enterprise artificial intelligence with Google Cloud through open agent collaboration. (sap.com, sap.com) For marketers, the immediate pitch is less about chatbots and more about software that can plan and adjust campaigns across channels using customer and business data from both systems. SAP said customer availability for the first marketing use cases is expected in the second half of 2026. (sap.com) Google framed the broader goal as an “agentic enterprise,” where artificial intelligence systems can act across business processes instead of answering one prompt at a time. In a separate April 22 post, Google said the joint architecture is meant to turn cloud modernization and data integration into a shared foundation for agents. (cloud.google.com, cloud.google.com) SAP’s architecture materials show the company is also working on agent interoperability rules so SAP agents can exchange context and actions with agents running on other clouds, including Google Cloud and Microsoft Azure. That gives this partnership a role in a larger contest over whose platforms become the control layer for enterprise artificial intelligence. (architecture.learning.sap.com, cloud.google.com) The next marker is whether customers move from pilots to production when those marketing agents arrive later in 2026. For now, SAP and Google are selling a simpler promise: keep the data in place, let the agents coordinate across systems, and cut the manual work between them. (sap.com, cloud.google.com)