SAP Autonomous Enterprise

SAP just put a number on the agentic enterprise: 200+ specialized AI agents, 50+ domain assistants, all running through a single platform.

At Sapphire 2026 in Orlando, SAP unveiled the Autonomous Enterprise, the most aggressive bet yet that ERP's next chapter is not "AI features bolted on" but AI agents running finance, supply chain, procurement, HR, and customer experience end-to-end.

Three things make this different from the AI roadmap announcements we have seen for the last 24 months.

First, the reasoning layer is Anthropic's Claude, embedded directly into the SAP Business AI Platform. SAP did not try to build its own frontier model. It plugged in the most capable one and focused on what only SAP can do, which is govern the business data and processes underneath.

Second, the SAP Knowledge Graph gives those agents a structured map of how a specific enterprise actually runs. That is the missing piece in most agentic AI pilots today, and it is the reason most enterprise AI projects never make it past the demo.

Third, the agents are designed to orchestrate across functions, not just automate single tasks. A procurement agent that cannot talk to a finance agent is not autonomy. It is automation with extra steps.

The strategic question for operators is no longer whether agents will run core business processes. It is who owns the platform they run on, and how much of your workflow data you are willing to hand to it.

What governance guardrails would you want in place before letting an AI agent close your books?