During SuiteConnect London 2026, Oracle NetSuite announced three new enhancements to its AI Connector service. But behind these announcements lie strategic choices that go beyond functionality. NetSuite decided to stay out of the model wars and avoid them. A conversation with Craig Sullivan and Patrick Pack reveals what that’s really like.
Last year, we asked Sullivan and Pack at SuiteConnect in London whether external AI models could also access NetSuite data. The answer back then was still very cautious, and even then it turned out to be very difficult to implement. This year, we asked the same question again, but the answer was radically different.
“It was a turning point, really just in the past year,” Sullivan admits. While the maturation of models like Claude and ChatGPT had an impact, it was the arrival of the Model Context Protocol that brought about the real change. “As soon as OpenAI adopted MCP, we jumped on it,” Sullivan says. “We were probably the first ERP system to embrace MCP as a way to connect AI to our core business systems.”
This is a surprising shift in perspective for companies that have traditionally wanted to apply AI primarily within their own platforms. The market is simply forcing it to open up. Organizations are already using Claude or ChatGPT separately from their ERP. The question is no longer whether these tools exist, but how do you make them work responsibly for your business data?
Three extensions with the same AI strategy
The announcements at SuiteConnect London are three extensions using essentially the same AI strategy. AI Connector Service Companion helps you ask the right questions of your data. Over 100 financial prompt templates tailored to NetSuite data structures, roles, and terminology. CFOs don’t need to be nimble engineers to ask meaningful questions. Templates are organized by business process and user role, from administrator to financial analyst.
Additionally, there are MCP apps (AI apps) that solve user interface (UX) issues. Instead of empty text boxes, give your external AI assistant the familiar NetSuite interface with filters, selections, and reports directly. For example, I saw a demo of Claude that allows you to generate dashboards and HTML pages with complete reports that you can immediately view within Claude. In ChatGPT, its visual layer works through native UI support. Gemini is not yet compatible, but will be supported as well as soon as possible. “If they support MCP as a client, that immediately helps,” Puck says. “We don’t block anyone.”
Our latest announcement is NetSuite Analytics Warehouse, which enables AI access to historical, analytical, and external data. This allows AI to perform analysis not only on today’s transaction data, but also on multi-year trends, Shopify data, or other systems connected through the warehouse.
Also read: NetSuite expands AI Connector service with MCP app
Knowledge of 43,000 customers in a rapid library
In theory, a companion is a library of prompt templates. But in reality it is fundamentally different. It is the expertise of experienced accountants translated into software. This makes it easy for people with limited financial knowledge to ask the right questions of their data.
When you ask Sullivan how NetSuite decides what’s included in these prompts, the answer is clear. “What we learn from industry experts and our 43,000 customers continues to grow,” he says. “Insights that come from implementation and conversations with customers that help shape the next generation of platforms.” Puck added, “This is actually a really exciting part of what we’re doing right now. There’s a lot of value in creating these prompts and best practices.”
Key market and industry knowledge is transformed into reusable prompts. Good managers know what questions to ask at the end of a quarter, how to interpret cash flow variances, and when reconciliation discrepancies are errors or timing issues. That knowledge is embedded in the prompt template. Now it’s available to everyone.
We ask Sullivan directly. Would you replace your employee with this? He chooses his words carefully. “I’ve been in this industry for a long time and I’ve talked to a lot of customers. It’s unusual for a customer to say that their employees have too little to do. There’s always more work than people.” His perspective is that AI will enable people to contribute to tasks that were previously untouched. Prompts do not replace thinking. They take it to a higher level.
Internal and external: combination, not selection
Logical follow-up question: When integrating an external AI model, can it also replace the internal model? The answer is no, and that’s a deliberate architectural choice. NetSuite’s unique AI assistant, Ask Oracle, is deeply embedded in the NetSuite architecture. External AI tools work alongside AI, not instead of it.
But what’s going on inside Ask Oracle is more interesting than you might think. NetSuite doesn’t use a single model for everything. “We use the right model for the right task,” Pack says. It sounds obvious, but there are underlying layers to choosing the right model.
We also asked about the development of a relational foundation model. These types of models are specifically designed for reading and analyzing financial data. You can make predictions based on smaller datasets without training. Pack responded, “This is a really interesting space. We’re monitoring it closely and there’s a lot of research going on internally, but I can’t talk about it yet.”
What he can tell you is that NetSuite builds its own machine learning models for specific use cases and combines them with LLM models. The direction is clear. Agent AI can also make predictions with the right support, context, and data. “There are a lot of really interesting use cases for combining agent AI and predictive AI to enable agents to predict outcomes, use inference capabilities to make recommendations, and take actions,” Puck says. Sullivan calls this an Oracle advantage. Through the OCI infrastructure, NetSuite has access to a variety of models, including LLM, predictive models, and proprietary ML models, which are seamlessly and transparently combined for users. Anyone typing in a text field will get the best answer, regardless of which model generated that answer.
Bottom line: the platform wins, not the model.
NetSuite’s AI strategy is as open as its platform. NetSuite aims to become a trusted, context-aware AI platform for financial processes by working on open standards (MCP) and powering its own AI with a combination of LLM, predictive models, and in-house ML. For them, it doesn’t matter which AI model a customer chooses to use.
The direction is clear. However, implementation is still a work in progress. OpenAI and Claude are up and running. It seems like it’s only a matter of time for Gemini. But given the pace of innovation in the world of AI, next year could be a very different conversation.
