“We’re like White Lotus,” Evan Goldberg grins. “We’ve got to have you hanging on the edge of your seat for the next episode.”
It’s a bold analogy for ERP software – but Goldberg, founder and EVP of Oracle NetSuite, says the company’s AI strategy is reshaping its future just as dramatically as the iconic HBO drama.
When Goldberg founded the cloud-based NetSuite in 1998, its edge was being accessible anywhere from any browser. It was a revolutionary idea for ERPs at the time. Today, he argues that NetSuite’s AI push gives it an even bigger edge than it had 30 years ago.
So how is NetSuite doing it differently?
First, by embedding AI into its software, not selling it as an add-on. That decision means it misses out on charging extra for AI but creates deep structural advantages, Goldberg says. Second, NetSuite is introducing AI agents to automate tasks traditionally done by finance staff. And third – most significantly – AI could create an interface that lets a novice use an ERP like a power user.
The implications for ERP design and user interfaces are huge. (For more on AI and interfaces, see the companion interview with head of AI and Technology, Brian Chess.) Goldberg even suggests that with a chat-like interface, ERPs could become simple enough for a corner-store florist to use – and stick with all the way to IPO. But doesn't that mean NetSuite could take on small business software, like Xero and QuickBooks?
In this wide-ranging interview, recorded at Sydney SuiteConnect, Goldberg lays out his vision for AI in ERP – and leaves with a cliffhanger on what is coming next.
The interview below has been edited for clarity and length. It was recorded in person at the SuiteConnect Sydney event in Sydney, Australia on 4 April, 2025.
Scale100: I just had a very interesting chat with [NetSuite head of AI] Brian Chess about the potential impact of AI, and generative AI especially, in ERP. But what’s your take? Would you ever trust generative AI to write financial records?
Goldberg: So today, we showed financial exception management. What that just does is surface things that the AI thinks are out of whack. And if it's right most of the time but occasionally it [classifies a transaction as suspicious when it’s] totally fine, you look it over and you're like, “No, that's right”.
Scale100: Will we ever trust it to write to records?
Goldberg: It's a really good question. It's kind of like self-driving. It's probably going to have to be 100 times better than humans – which is not that hard because humans are bad drivers. So it won't be that hard for it to be better, but it'll have to be a lot better before people trust it.
But we have human beings that are entering this financial data and they make mistakes. If we can make it significantly more accurate than your everyday trainee, or you look over the lifetime of an employee and compare how many mistakes they make versus the AI, that's the calculation that should be made.
That's been the history of a lot of these systems. You start by giving them not very much autonomy, and then over time you get more comfortable with it. I think that's going to happen.
Scale100: I love that analogy with self-driving. I've been reading a lot about Elon Musk’s promises about self-driving and this key statistic, the number of miles between disengagements…
Goldberg: I live in San Francisco and we have these Waymos literally everywhere. North Beach, where I live, is a very popular spot for dinner. On a Friday night every car is a Waymo. It's crazy. They're not getting in accidents, people are getting where they're going, and people love them. I don't have to have a conversation [with an Uber driver].
Five years ago, people would say, “No way, I could never trust that”. But thousands of people are getting in these Waymos and getting where they want to go. It's pretty incredible.
Scale100: So everyone's talking about agents. Do you think that we will see, say in a five-year horizon, ERPs creating finance teams using agents?
Goldberg: We talked today about how finance is becoming more strategic because companies are becoming more data driven. They want to leverage data to make really good decisions. The finance team is a really great asset for that [purpose] because they have those skill sets and they [can] make a difference not just on your P&L statement but in operations.
These people in finance already are so overwhelmed with all the things they could do. They're just going to be given more responsibilities. It’s the story of life. “Here's something else you can do. Try to fit it into your already very busy day.”
How could we help them help the business run better? AI is going to be that assistant that helps them get a lot more done. They trust that the AI is going to do it right so they can apply their human judgment and get more involved in all parts of the organisation.
So when I think about agents, I think about them as tools for people in finance and all over the organisation to do more when they're not going to be given any more to do it with.
Scale100: I'm doing some work helping a mid-size construction company with their innovation strategy. The finance team deals with sub-contractors who send in the broadest range of things that you could call an invoice, and they've got to pull the relevant parts out into a spreadsheet.
Goldberg: That's a great example of where AI can probably do it with 98% accuracy. That's the type of function it's going to get really, really good at. The systems that I've been using tie together lots of different technologies. They don't necessarily use a large language model. They might have an OCR [engine] which the software can operate and converse with, basically how a human user would use it, but behind the scenes.
Scale100: Are we talking about agents as just another piece of software here, or are we saying it's a piece of software that has the ability to converse?
Goldberg: We talked about agents today and the fact that you can converse with them in natural language is one of their most important attributes. And that they know the context of your business. So they know who your vendors are, and know what kind of things you see in your invoices. They get that and then they learn. All those attributes make a really good assistant that is going to do stuff for you so you can level up and work on more strategic tasks for the company.
Scale100: Brian [Chess] talked about conversational interfaces for business users like sales managers and business owners. Is NetSuite interested in going down market? Because you could have a florist and behind the curtain they could be using the world's biggest machine…
Goldberg: We are definitely interested in seeing how we can make NetSuite so easy to use that implementing it in a very small company becomes more viable.
Scale100: Do you think then we'll have a big consolidation? I mean, you won't need a Xero or QuickBooks if you can just go with –
Goldberg: How cool would that be? “The last ERP system you'll ever have to use.” When that florist goes public, they're still on NetSuite.
Scale100: But why not?
Goldberg: That is definitely a vision that we have.
Scale100: So are you redrawing the TAMs? Is this real?
Goldberg: More to come.
Scale100: Why?
Goldberg: Well, you can't get it all in one session. We've got to keep you coming back.
Scale100: Well, it's not the first time we have had this chat.
Goldberg: I know, but we're like White Lotus, we’ve got to have you hanging on the edge of your seat for the next episode.
Scale100: Ok – so just thinking about opportunities and challenges for NetSuite globally, is there anything else that you are talking about or is AI the main thing?
Goldberg: I would definitely say it's the main thing. It’s like [how important] the internet was at the beginning of our company. Our big differentiator then was that you could access it anytime, anywhere. And our other big differentiator was that it was all the tools you needed in one place. So I would say that our differentiator over the next few years is going to be, “We have the best AI for a business system”. Because that gets to the heart of what these systems should do – advise and assist. AI is going to make them so much better at advising and assisting.
Scale100: So how do you define that it is the best system, and how do you deliver that?
Goldberg: Hopefully it'll be our customers that will be able to articulate that. We're going to listen to our customers and do a lot of experimentation and things like financial exception management. We have 41,000 customers, so we have a really great test bed. And they're going to come back and say, here's what it does. As with the self-driving example, if it took a wrong turn, we'll fix that. The fact that we have this large, very diverse customer base across so many different industries and regions – that's how I think we're going to end up being the best.
Scale100: But surely SAP and Microsoft are going to be saying something similar.
Goldberg: Everybody has their strengths, and in industries that they're particularly strong they'll probably be really good at it. We just happen to have a cloud-based customer base that I think is unparalleled. And the other thing is the resources of Oracle. Because we're definitely going to be comparing notes with our colleagues that do Oracle Fusion. And that's another several thousand companies that are very, very big companies, the size that maybe many of our customers aspire to be. So that's the richest data set in business software bar none, including Microsoft, because it encompasses so many of the functions of these companies. It has that breadth. It has depth. It's big. I think Oracle has something that's unparalleled in that regard.
Scale100: Is it a "He who has the most data wins" situation?
Goldberg: It's certainly helpful. I mean, I would say it's [Oracle’s] to lose for sure. We have to obviously take advantage of that rich data set that we have, and that's what we're fully intending to do.
Scale100: Which specific industries will benefit the most from building AI into the ERP?
Goldberg: I was just talking to a customer right now who said, we are a 50-year-old business and we have every type of user you can imagine, with every level of familiarity with technology. It's really important to us that all of them are able to get the most out of the system. There's so much opportunity for a company like that.
So it's not necessarily one particular industry or the other. Where you have companies that don't have as many IT resources – they're not necessarily tech companies – and they have a variety of users, I think they're going to get this dramatic leveling up of what they can get out of the system.
Scale100: One of the interesting concepts Brian talked about was opening the ability to code to finance teams. The finance people I know probably wouldn't feel that comfortable with technology, but it's an interesting concept.
Goldberg: Although they tend to do magic with spreadsheets.
Scale100: That's true.
Goldberg: You know, a lot of that [spreadsheet work] is really programming, in some sense. If you can harness somebody that can actually write any code in any language, it's really powerful.
I think we're going to find that there's a lot of people that just take this stuff and run with it and do incredible things. We already have customers that are doing this. Their finance teams are using gen AI to write scripts. A lot of the things that they used to do by combining spreadsheets – the AI can just do it for them when they ask for it in everyday English. It’s incredibly cool.
Scale100: So how does that change the characteristics for how you hire people for your finance team?
Goldberg: At the top level they have to have that analytical mindset that helps them deal with numbers. I think this is just a new tool for them to adopt. You don't necessarily have to hire someone that has vast experience with Excel. You can get down to that analytical mindset and other softer, qualitative attributes rather than just pure skill sets, because a lot of these skills you can gain through the use of large language models.
Scale100: I'm just imagining that you’d want people more self driven, willing to experiment… These are characteristics that not all accountants would be as comfortable displaying.
Goldberg: No, that's true. My experience is that the best finance people are creative because they can figure out how to measure and control something that gets you the business outcome you want, and in a way that's repeatable. That's a different type of creativity. What's interesting is the mix of attributes that you want in the future. I still think those people will be able to get an enormous amount out of AI, even if they don't have all these attributes.
I really think it's going to help everybody. I subscribe to the notion that, all the way back to the steam engine and John Henry, society has adapted to technology and people have had lots to do. I just think there's so much to do in companies that everybody's going to find a place to be, and AI is just going to make them happier and more productive. But that's optimistic.
Scale100: Yes, particularly if you've got a company that stands to benefit enormously from AI.
Goldberg: I mean, yes and no. Look, without AI we have great business. We're just trying to do the best for our customers, and that's what it's all about. And if it's helpful for our customers, this is a much better version of NetSuite.
Scale100: It's interesting to hear you say that you feel like this is a watershed moment, like when you founded NetSuite, and yet NetSuite is here for the second time, a second evolution.
Goldberg: I have Larry Ellison as a mentor, who made the transition to software-as-a-service incredibly well, and now has made the transition to infrastructure-as-a-service. So starting as a database company and now becoming, 50 years later, so central to all these different trends, I feel like it's very doable.
Scale100: Why can NetSuite pull it off when others have failed?
Goldberg: Well, it really is connected to what I said about Larry. I think when the founders are there, and they still have that energy that they had when they found the company, you can translate that [to the next revolution].
I left Oracle after three years when I just got out of university to be in a doctoral program for AI. So I've been really interested in it all along. I think if you have founders that are still involved in the business and have that same kind of passion that they did at the beginning that can be a fuel. It sounds kind of self-serving, but I'm really more talking about Larry than I'm talking about me.
Image credit: Montage uses photos supplied by Oracle NetSuite from the SuiteConnect Sydney event in Sydney on 4 April 2025, and a frame screenshot from HBO White Lotus.
