There is a moment in every major technology cycle when the market’s imagination runs ahead of its understanding, and right now artificial intelligence has pushed us into exactly that kind of moment. Every week seems to bring another prediction that AI will hollow out entire categories of software, replace armies of knowledge workers, and rearrange the economic order of the enterprise. In the middle of this noise, ServiceNow President and CFO Gina Mastantuono is offering a calm but firm counterargument. The idea that AI is somehow “disrupting” enterprise software, she says, is simply false. That is not a small statement. ServiceNow is one of the most closely watched companies in the software industry, and Mastantuono is one of the most senior finance executives in tech, so when she pushes back on the disruption narrative, it matters. But the timing of her remarks is also significant. Only a day before her interview, ServiceNow shares fell after the company reported quarterly results that, on the surface, looked solid. The fact that investors sold off the stock anyway says a great deal about the mood of the market. The same AI that is supposed to be an engine of innovation has become, in the minds of many traders, a reason to doubt the future of traditional enterprise platforms. Mastantuono’s message, delivered with the calm of someone who has seen hype cycles before, is that this fear is misplaced—but so is the complacency that would ignore the real risks AI brings.
The market drop that preceded her comments was not necessarily a sign of corporate trouble. ServiceNow reported results that included strong subscription revenue and continued growth, and the company’s forward-looking guidance was, by most accounts, respectable. But in the current environment, respectable no longer feels sufficient. Investors are looking for evidence that AI is not cannibalizing existing software sales, and they are trying to figure out which companies will be the winners and which will be the losers in a world where AI can write code, automate workflows, generate reports, and answer customer questions. The worry is that if AI can do all those things, then why would a company need to buy a large, layered software platform? Why not just buy an AI model and let it figure things out? That concern has weighed on the entire enterprise software sector, and ServiceNow has found itself in the crosshairs because it sits at the center of so many business workflows. Mastantuono’s response is not dismissive. She acknowledges the anxiety and the genuine questions behind it. But she is also clear that the mental model underlying the fear is wrong. Enterprise software is not being replaced by AI; it is becoming the infrastructure through which AI can actually deliver value. The two are not competitors. They are partners, and any business that thinks it can deploy AI without a strong underlying software and governance structure is likely heading for trouble.
What does Mastantuono mean when she says that every business and every industry needs a “harness” for AI? The image is a useful one, because a harness is not a cage. It does not stop the animal from moving. In fact, it gives the animal direction and allows its power to be put to useful work. Without a harness, a strong horse can run wild, injure itself, or knock over the wagon it was supposed to pull. With one, it can carry an enormous load across difficult terrain. Mastantuono is making the same point about AI. AI has enormous power. It can generate text, summarize information, detect patterns, predict outcomes, and automate tasks that used to require hours of human effort. But that power, left unstructured, is dangerous. It can produce confident but incorrect answers. It can reinforce bias. It can make decisions based on incomplete context. It can surprise a company with legal, reputational, and operational consequences that no one anticipated. The harness, in her view, is the set of controls, workflows, governance, and human oversight that turn AI from a curious experiment into a reliable business tool. That harness does not emerge from nowhere. It is built into the enterprise software platforms that companies already use, provided those platforms are modern enough to understand and manage AI-powered processes.
Seen this way, the real conversation about AI should not be about disruption. It should be about integration. Enterprise software has always been the connective tissue of an organization. It keeps the sales team in sync with the finance team. It makes sure a customer support ticket reaches the right person at the right time. It tracks budgets, approvals, risks, and service levels. It is the memory of the organization, preserving the rules and processes that make large companies manageable. AI layers on top of that. It can suggest the next best action for a customer service agent. It can flag procurement anomalies before they become fraud. It can route a complex HR case to the right specialist with the right context. But to do any of those things well, AI needs to understand the organization’s structure, its data, its policies, and its history. That understanding does not exist in a vacuum. It lives in the enterprise software stack. Mastantuono’s argument is that AI makes enterprise software more valuable, not less. The more capable the AI, the more important it becomes to have a system that ties outputs to actual workflows, approvals, and accountability. Without that system, AI is a clever parlor trick. With it, AI becomes a transformative force. Investors who see AI as a threat to software are, in her view, missing the real story: the companies that will dominate the next few years are the ones that learn to pair their software platforms with carefully governed AI capabilities.
At the same time, Mastantuono is not arguing that businesses can relax and simply trust AI to sort itself out. Her call for safeguards is a practical one, not a philosophical one. She is urging business leaders to think about AI deployment the way they would think about any other major operational change. That means knowing exactly what the AI is being used for, what data it is drawing from, and what risks those data carry. It means establishing clear ownership for AI-driven decisions, so that when something goes wrong, there is no question about who is accountable. It means putting humans in the loop for high-stakes outcomes, not because humans are always right, but because human judgment is still needed when the stakes are high and the context is complex. It means testing AI outputs continuously, not just at launch, and being willing to pull a model back or retrain it when its behavior drifts. It means being transparent with customers, employees, and regulators about the limitations of AI, and building mechanisms that allow people to challenge or appeal an algorithmic decision. These are not exotic requirements. They are the same kinds of reliability, auditability, and control capabilities that enterprise software has provided for decades. The difference is that the pace of AI change is so fast that the safeguards must be designed in from the beginning, rather than bolted on after a crisis. Mastantuono’s language is reassuring, but it is also demanding. She is saying, in effect, that AI is not something to fear and not something to ignore. It is something to manage, and managing it properly is a job for every leader in every industry.
Perhaps the most humane part of Mastantuono’s message is that she does not ask anyone to pretend the market’s anxiety is baseless. The share price decline after ServiceNow’s earnings report is a reminder that in the current climate, perception can move against a company even when its fundamentals are stable. That is a hard truth for any CFO to swallow, and her willingness to speak openly about it is a sign of emotional and strategic maturity. She is not telling investors to be blindly optimistic. She is telling them to be more accurate in how they think about AI. Disruption is a myth in the sense that AI will not simply rip up the enterprise software model and throw it away. But disruption is real in the sense that AI is changing the expectations of what software should do, how quickly it should adapt, and how deeply it should support decision-making. The companies that survive and thrive will be those that accept both truths at once. They will keep building and improving their software platforms, while also folding AI into everything those platforms do. They will treat AI as an amplifier, not an intruder. They will build the harness before they need it, rather than searching for one in the middle of a storm. And they will understand that the fear surrounding AI is not just a market distraction. It is a signal that leaders need to communicate more clearly, govern more carefully, and prove—through action, not just words—that AI can be safe, useful, and humane. In that way, Mastantuono’s comments are less a defense of one company’s stock price and more an invitation to think about the future with both discipline and calm. The software industry has always been shaped by cycles of fear and enthusiasm. What matters is the work done between those cycles: the difficult, unglamorous work of building systems that people can rely on. That work, more than any single technology, is what creates lasting value. And in the era of AI, that work demands a harness.