Oracle NetSuite counters the AI investor narrative with its own AI long game - inside NetSuite's AI news with Evan Goldberg

You can't make three clicks on LinkedIn without reading about how SaaS is in trouble. Supposedly, the potent-but-flawed tech we call "AI" today will eat everything in its path. 

Nevermind that savvy SaaS players have been investing heavily in embedding LLM features, while curbing the worst tendencies of LLMs with proper data context. 

You've probably heard my rant on that - but what does Evan Goldberg think? Now's the perfect day to ask. On Wednesday the 11th, timed with SuiteConnect NYC,  Oracle NetSuite released a slew of AI news, including: 

• NetSuite Announces AI Innovations to Help Businesses Increase Efficiency (including new automated close, reconciliation, and pricing capabilities)
• NetSuite Helps Organizations Reduce Application Integration Challenges with New AI-Powered Platform (officially called the NetSuite Integration Platform, intended to help customers "automate, orchestrate, and manage cross-application workflows.")

How can cloud ERP vendors stay relevant? My checklist for AI readiness

In our sometimes-fickle market economy, SaaS is at a "prove it to me" crossroads. As I see it, cloud ERP vendors need to accomplish the following things: 

• Demonstrate that their existing customers are already achieving significant value on a modern platform, not overspending on outdated, inflexible pseudo-cloud applications. Impactful customer use cases matter!
• Show how prior investments in deterministic automation, role-based security, localization and industry compliance are not easily replaceable, but highly relevant to applying agentic AI.
• Make a convincing case that the data and process discipline imposed by running on modern ERP - including a new level of data visibility - are the best building blocks for enterprise-grade AI, not a detour.
• Earn customers' AI trust by managing the sophisticated AI architectures that deliver better results, including LLM model management/testing, context engineering, managing external tool calls, etc. (This is especially important for SMB customers, most of which don't have the tech resources to contend with issues like model drift, RAG chunking optimization, LLM evaluation, explainability, minimizing biased/inaccurate output, and other fun essentials).

Inside NetSuite's AI for finance news

How does NetSuite score on these items? Let's have a look at Goldberg's comments. Goldberg, Founder and EVP for Oracle NetSuite, has seen plenty of hype cycles rise and fall on his watch. So how does he perceive NetSuite's AI play? How does this news fit into NetSuite's long game? As Goldberg told me: 

If you talk to our customers and companies like theirs, AI has obviously shot up in their list of concerns. Especially for our customers that don't have entire IT departments where they can say, 'Okay, now go off and research AI and figure out how we're going to integrate it into all our processes'. 

The ability to rely on the fact that we're building it into every corner of NetSuite over time, is a way that they can be assured that they're not missing the AI boat, but we're doing it in a way that is really intrinsic to their processes, just a natural extension of the things that they're already doing - taking some of the things that they're used to doing manually, and having them be automated, while keeping a human in the loop. That's sort of the general principle that we're going with.

If you want your customers to trust AI, nailing down the security of data access is not negotiable. Goldberg: 

Just building it deeply into the system, but making sure that they can trust it, that it uses all the same sort of protections and permissions that the system that they have already set up the system to do, because they know that you know that they have a variety of different types of data that you know they want some people to see and other people not to see. And AI has to respect all of that. 

As I see it, if you want to deliver AI to businesses that has real impact, you better solve something for a role or an industry - don't just give me a generic "write me an email" productivity tool. Goldberg elaborated on the NetSuite AI news for finance with this in mind: 

 I think that the AI that we're giving to finance teams to help them close faster and eliminate manual processes like the AI bank matching capability, the intelligent payment runs. Most people don't think of payables as something that can be a strategic asset, but if you can use AI to help you decide what order you're going to pay things, and when to satisfy more strategic considerations, making sure your critical suppliers are getting paid on time, so that they keep delivering for your best customers, making sure you're taking advantage of early payment discounts. 

That's the type of stuff you manually had to go through and in the past and figure it out, but now, this is how AI sort of works everywhere. When it's working best is: you give it a high level objective, and then it goes and does things under the covers. It tells you what it did, so you can have that sort of transparency and trust. And it goes ahead and does them behind the scenes. 

On job impact, pricing - and how probabilistic and deterministic tools fit together

I expect vendors to have a narrative on the jobs impact - hopefully as open as possible, whether I agree with it or not. Goldberg beat me to that: 

People talk about, 'Well, what's going to happen to these jobs?' What's going to happen to them is: everybody in every profession has a list of urgent stuff, and a list of important stuff. Ror finance teams, obviously, the urgent is making sure that books are correct, and ideally, making them correct, not just at the end of the period, but all the time, so that the management of the organization can always have a clear and up-to-date view of what's happening in the business. So that's sort of the urgent things.

If they can get AI to make those thing easier and faster and do work behind the scenes, then they get to their important list, which is: some of the more strategic things that they're going to do to help the business. That's the theme of everything we're doing: automation gives you time back, and then you can use the insight and adaptability of the platform, the agility of the platform, to continually improve both NetSuite within your business and of course, the operation of your business overall.

Whenever I hear executives talk up the virtues of agentic AI automation, I want to ask them about their deterministic workflows. I want to hear: how are these new agentic workflows different? And: are we still taking advantage of the deterministic workflows where they are needed? Because even the most well-design LLM agent is not deterministic, that's just a reality of this tech. Goldberg responded: 

We're on the same journey as all our customers, and learning how to best use this AI technology. But there are situations where we really can kind of combine the best of both worlds, right? The Suite Cloud Developer Assistant uses AI to develop a NetSuite extension, but the NetSuite extension then uses SuiteScript or SuiteFlow, these deterministic technologies, to actually implement it. So if it says it's going to do something, it's going to do it exactly as it says. Combining the well-known automation and extensibility capabilities of NetSuite with AI to help you use them quicker and more efficiently may be sort of the sweet spot, in that tension between probabilistic and deterministic.

NetSuite has the advantage of providing SMB customers with something much closer to a single source of truth than large enterprises can usually get to (large enterprises tend to have a dizzying array of software footprints, and they are one big merger of acquisition away from acquiring another). But that doesn't mean there aren't integration challenges. NetSuite thinks it can impact this with AI, via the NetSuite Integration Platform announcement: 

People use NetSuite as their system of record, and it replaces, in many cases, multiple systems. But every business has multiple systems, and they're always trying to get them to work together. It's one of the hardest problems, and AI can really help. One of the problems doing integration is the the connections can be brittle. They can break. You need specialists that are able to fix them when that happens. But when AI builds these connections, it can be doing that, and you can rely on the AI going forward to help you maintain these integrations, and change them.

One final component is pricing. NetSuite has generally followed the same sensible pricing approach as Oracle itself - not charging for most embedded AI features, charging a premium for some value-added services. It's a far more sensible approach than most of NetSuite's competitors, and one that should help greatly with adoption. But there are still open questions, such as: does consumption pricing make more sense with AI, at least with some roles? And what about heavy adopting customers (or teams) that drive up inference costs with big usage spikes? Goldberg: 

It's a great question, and I don't think anybody totally understands how it's going to evolve. For example, in some cases, I think AI's ability to use the system more effectively with little or no training is actually going to increase your user count. More people will want to use NetSuite when they see how easy it is with AI. In the long term, you're absolutely right that people may figure out how to use AI in different ways that can change the balance of value. 

Our philosophy going in right now is first, we're just going to make sure everybody has AI. Because I've said many times, these systems are going to be nothing without AI. There's not going to be an AI-less NetSuite. And so we're going in with this notion that we're going to open the system for all users to use AI...  It's not an imminent threat to our costs or revenues, and so we're taking an approach of letting everybody in for free.

My take - AI customer adoption is everything now

Of course, even bigger AI adoption stories are ahead, as we watch the next-gen NetSuite Next interface unfold. First announced in October 2025 at SuiteWorld, NetSuite Next will put an agentic interface as a guiding feature of NetSuite's applications, via the Ask Oracle enterprise assistant tech. (Goldberg noted that as customers adopt NetSuite Next, they'll be able to do it alongside their existing systems on a per-user basis, and set their own adoption pace). 

As you can see from my initial list, NetSuite meets most of my criteria for the kind of SaaS that should gain from AI, not be undermined by it. One criteria that I left out: LLMs are a big-data-type-of-AI in the extreme, and benefit from economies of scale. NetSuite has that advantage via its parent company, as demonstrated by their use of "Ask Oracle" tech in NetSuite Next (last fall, Brian Sommer laid out the advantages NetSuite has gained from Oracle's global reach, including extensive use of Oracle Cloud Infrastructure). Speaking of Sommer, he was on the ground in New York City this week; look for his event roundup in the coming days on these pages...

What we'll need to track most this year is customer adoption. The gap between the pace of development innovation - versus what customers can absorb - is about as wide as it's ever been. That's not necessarily a negative, as long as vendors help customers make sense of what to do next - and identify what will have the most impact per adoption phase. 

Here, partners can help - but partners are also potentially threatened by some of the AI services vendors like NetSuite can now provide. Goldberg sees this as a positive for NetSuite's partners, however, for the same reason he provided for finance teams. Now, Goldberg says, NetSuite's partners can let AI services handle some of the "urgent" tasks, and then partners can focus on delivering vertical expertise - and building IP for customers. 

If customers can integrate disparate systems more easily with NetSuite's Integration Plaform, that will be a notable win. However, I'll need to hear more from customers on just how effective AI can be for this type of integration. Reducing the burden is one thing; eliminating it entirely will require that my questioning ears hear some customer use cases in action. But I love the ambition... Going after customers' landscape pain points is an undernourished enterprise AI plotline.

Beyond customer adoption, I can think of a slew of other challenges, not the least of which are the fickle/volatile markets and their impact on capital - and customer sentiment. NetSuite's long game looks like the right one to me; it will be fascinating to see if NetSuite has the runway of product wisdom to execute on it while so many disruptions, perceived and imagined, weigh on software buyers and vendors. 

The same goes for automation. My view is that savvy ERP customers should already have plenty of useful automations in place, prior to any agentic AI usage. Goldberg talks in terms of moving teams from "urgent" work to "important." This is similar to my long-time contention: the true hallmark of "intelligent" software is flipping our roles from 80% admin/20% strategy to the reverse. That doesn't happen overnight. 

Most companies haven't come close, though modern SaaS with low-code, workflow automation, and real-time analytics is the best step I've seen. I do believe agentic AI can take us further, if we adopt it properly, but I also think the dangers of so-called AI "workslop" are real too, not to mention stressed out/sped up workers in fear of job displacement or productivity scrutiny. (I don't see NetSuite contributing to this workslop with the type of gen AI they are utilizing, but that doesn't mean NetSuite customers won't deal with that from other forms of internal gen AI usage). 

Here, I give Goldberg credit: not once in our talk did I hear the usual dreaded "digital worker/teammate" buzzwords. Instead, I heard an emphasis on enhancing the strategic impact of human work, with (human) supervision of AI workflows where needed. True, NetSuite can't truly control how its own customers apply these tools internally, whether they utilize them primarily in the service of headcount reduction, etc. 

But I'm a strong believer in setting the right narrative frame - and what I heard from Goldberg is a valid and worthwhile framing for 2026. Let's see where this goes from here.