top of page
Search

NetSuite announces MCP based agnostic AI Connector Service

ree

NetSuite is introducing a protocol-driven integration service called AI Connector Service that allows companies to “bring their own AI” into the ERP according to an announcement by Netsuite. Rather than locking customers into one model, they’re supporting the Model Context Protocol (MCP) — a standard for securely connecting external AI agents to enterprise systems.


What this means in practice:

  • Customers can choose which AI models to use.

  • Developers define the permissions and access levels.

  • Partners can create new “AI-powered SuiteApps” to extend ERP functionality.

  • Most importantly, it aligns with NetSuite’s existing extensibility model, reducing the need for workarounds.


A few weeks ago, I wrote about NetSuite’s AI showcase and stressed a simple but important point: AI is only as good as the data you feed it. “Garbage in, garbage out” has never been more true, and I argued that NetSuite’s strength lies in its ability to centralize and normalize data across finance, operations, CRM, and HR. That foundation is what makes ERP a natural home for AI. At the time, I said NetSuite was still in the early stages of its AI journey, with more impactful deployments ahead. I believe this new Connector is one of those steps. It makes AI integration with NetSuite more robust, secure, and immediate for companies that are ready.


In a market that’s moving this fast, being AI agnostic seems to be the smart play. It gives customers flexibility, protects them from vendor lock-in, and allows them to adapt as the AI landscape evolves.


More importantly, this approach plays directly to NetSuite’s strengths. AI on its own is not a silver bullet — it needs clean, connected, normalized data to deliver real value. That’s exactly what ERP provides when implemented and managed well. With the AI Connector, NetSuite is effectively saying: we’ll give you the governance and structure — you bring the intelligence you trust.


As exciting as this is, finance executives should still proceed with care. AI has inherent limitations — hallucinations, inconsistent results, and unpredictable behaviors. My advice: test use cases in a sandbox environment before rolling them out in production. With AI, the principle remains the same: trust, but verify.

bottom of page