Over the past few months, I've talked with dozens of companies that run on NetSuite about what they're doing with AI. I went into those conversations expecting to spend most of the time explaining the basics. That's how it went a year ago, and things have changed quite a bit since then.
Most of the people I talk with now have been using ChatGPT or Claude, and many of them are now using these tools every day. They understand what the tools are good at and where they tend to go wrong. Many of them have already tried a project or two, and in a lot of cases those projects didn't work out. But they can usually explain why that happened, and that alone tells me how far they've come.
So for the most part, these companies no longer need to be convinced that AI is worth their attention. They already know that it is.
Where they're struggling is in two places.
Some of them still aren't sure what they want to do with AI. They know that they should be doing something, and they've probably been asked to "do something with AI," but they haven't connected that to a specific problem in their business. I don't worry too much about this group. Because if you talk with them and walk through how work gets done in their company, then the problems worth solving usually become obvious pretty quickly.
The second group is larger, and I think that it's the more important one. These companies know exactly what they want. They want to resolve fulfillment exceptions faster. They want the collections team to spend less time figuring out who to call. They want a reliable margin-by-customer report without waiting days for someone to put it together. They've identified the goal, and they understand AI well enough to know that it should be able to help.
What they can't figure out is how to apply it. And that's where I think that the opportunity is for those of us who work in the NetSuite space.
Here's what I mean. Take the collections example. There are several ways that you could apply AI to it. You could use a saved search to give a model the data it needs to produce a prioritized call list each morning. You could write a scheduled script that scores each customer and stores the result in a custom field, so that the dashboards people already use become more useful. You could build something outside of NetSuite that pulls data through SuiteTalk or SuiteQL and sends recommendations to wherever the team already communicates. Or you could use a workflow that only flags the accounts that need attention and leaves everything else alone.
Those are all reasonable options. But they have very different costs, and they require different things from the people who'll maintain them. They also fail in different ways. Choosing the right one depends on how the account has been customized, on which fields are really being used (and how), on who's going to support it a year from now, and on what happens when the model makes a mistake.
Every one of those questions is a NetSuite question. They're the kinds of things that NetSuite administrators, developers, and consultants deal with every day. The AI part of the decision is relatively small. The NetSuite part is most of it.
I've seen what happens when someone without that background makes the call. They tend to choose the option that looks the most impressive in a demo, because they don't know how much customization a typical account has accumulated or what that does to a plan.
I have a lot of respect for the AI specialists doing this work, and I want to be fair to them. But the plans they produce often assume they're working with a NetSuite account that's much simpler than the one the customer really has.
If you already know NetSuite well, then you have most of what this work requires. The AI side is the smaller piece, and it can be learned by experimenting with the tools against real NetSuite data and paying attention to what works. That part goes fairly quickly. The NetSuite knowledge is what takes years to build, and there's no way to rush it.
I also think that NetSuite is an unusually good platform for this kind of work. The data is structured. The business processes are already defined in the system. And there are well-established ways to read from it and write back to it. Many mid-market companies that are trying to use AI don't have a clear picture of their own operations. NetSuite customers do, and the people who know their accounts know exactly where to look.
The most valuable work in these projects now happens before anything gets built. Someone needs to understand the process, understand how the company has configured NetSuite to support it, and then decide where AI fits and what it should be allowed to do on its own. In my experience, companies are willing to pay for that, but they're having a hard time finding people who can do it.
The people who are best positioned to help are the ones who already know NetSuite. If that's you, then I think that there's more opportunity in front of you than you might realize.