Share with your CIO
Infor is positioning its ERP platform as the open alternative to SAP’s increasingly controlled AI architecture, and the contrast is sharpening fast. Through 6,000 published APIs and a data lake that streams in real time to Databricks or Snowflake environments, Infor lets third-party AI agents, including Microsoft Copilot, talk directly to its own ERP-native agents without requiring SAP-style architectural approval. A healthcare customer example shows agents handling cash distribution math in minutes rather than hours via spreadsheet.
What this means for your business
If your ERP vendor sits at the center of your agentic AI roadmap, the API governance model it chooses will determine whether your AI investments compound or stall. SAP’s April policy requiring third-party agents to operate inside SAP-approved architectures drew sharp criticism precisely because it puts the vendor, not the customer, in control of the orchestration layer. Infor is betting the opposite posture wins. Which side of that bet matters most to you depends on how much of your AI stack you’ve already built outside your ERP vendor’s ecosystem.
The deeper issue isn’t openness as a principle. It’s that ERP systems hold the transactional context, the purchase orders, payment terms, supplier relationships, that AI agents need to act rather than just advise. When a vendor controls how agents access that context, it controls the ceiling on every automation you can build. Infor’s “let your agents talk to ours” framing sounds collaborative, but it’s also a competitive moat: if Infor’s pre-built agents already cover the core workflows, the incentive to build outside the platform drops, which is exactly what they want. The openness and the lock-in live in the same architecture.
The vendor most likely to win this round isn’t the one with the most APIs. It’s the one whose agent-to-agent communication standard becomes the default that others have to integrate against. Model Context Protocol is the emerging candidate for that standard, and Infor’s own position on it matters more than its API count. If you’re renewing an ERP contract or evaluating a platform extension in the next 18 months, the question to pressure-test isn’t “how many APIs do you publish?” but “what happens to my third-party AI investments if I need to switch you out?”
Concept deep-dive: Agent-to-agent communication
Agentic AI systems don’t work in isolation. They complete complex tasks by delegating subtasks to other specialized agents, the way a project manager routes work to subject-matter experts. Agent-to-agent communication is the protocol layer that lets, say, a Microsoft Copilot assistant hand off a payment optimization request to an Infor finance agent and receive a structured result back. Without an open, standardized handshake between agents, every cross-platform automation requires custom integration work, which is exactly the friction ERP vendors are now competing to eliminate.
Based on reporting from How open ERP APIs and data speed up agentic AI integration, deployment, originally published 2026-10-17 11:42:00.
