Google’s Gemini models are coming to Oracle’s enterprise applications, and Oracle says it plans to use them for embedded AI inside NetSuite.
The key line to focus on comes from Evan Goldberg: “choosing the right model for the right use case is critical to helping them get more value from AI.”
Choosing. That one word is the crux of the story in my mind.
For years, the model was whatever your vendor shipped, and nobody in your building thought about it. Model selection is now a real decision with cost, accuracy, and data handling implications. In Oracle’s AI Agent Studio, customers and partners pick.
In NetSuite, Oracle is picking per use case for now. Either way, the questions land on your desk.
Which model touched that transaction.
What it cost to run.
Who signed off on it seeing your financial data.
What you do when a cheaper, faster model handles part of your close, and the numbers start to drift.
If you are implementing NetSuite this year, or turning on AI in one you already run, somebody owns those answers.
Not your vendor. Not your implementation partner three months after go-live.
If I put that question to ten finance leaders today, I am almost certain the answers would range from: “IT, I think,” to “our controller has been playing with it,” to a very honest “I have no idea whatsoever.”
That is not a knock on anyone. It is a job that did not exist eighteen months ago and has not been assigned yet.
Google’s Gemini models are coming to Oracle’s enterprise applications, and Oracle says it plans to use them for embedded AI inside NetSuite.
The key line to focus on comes from Evan Goldberg: “choosing the right model for the right use case is critical to helping them get more value from AI.”
Choosing. That one word is the crux of the story in my mind.
For years, the model was whatever your vendor shipped, and nobody in your building thought about it. Model selection is now a real decision with cost, accuracy, and data handling implications. In Oracle’s AI Agent Studio, customers and partners pick.
In NetSuite, Oracle is picking per use case for now. Either way, the questions land on your desk.
Which model touched that transaction.
What it cost to run.
Who signed off on it seeing your financial data.
What you do when a cheaper, faster model handles part of your close, and the numbers start to drift.
If you are implementing NetSuite this year, or turning on AI in one you already run, somebody owns those answers.
Not your vendor. Not your implementation partner three months after go-live.
If I put that question to ten finance leaders today, I am almost certain the answers would range from: “IT, I think,” to “our controller has been playing with it,” to a very honest “I have no idea whatsoever.”
That is not a knock on anyone. It is a job that did not exist eighteen months ago and has not been assigned yet.
So, who is that person on your org chart today?
Source: Oracle press release, July 30, 2026.
https://lnkd.in/dSB2-Nx9
hashtag#NetSuite hashtag#ERP hashtag#ArtificialIntelligence hashtag#CFO hashtag#Oracle