Will there be a time when enterprises ask AI for an app instead of buying SaaS?
Will there be a moment when, instead of comparing SaaS vendors, pricing tiers, features, and licenses, a company that needs a new app just asks an AI to build it?
On one side, ServiceNow’s Bill McDermott says he is not worried that customers will replace what took them 20 years to build. On the other, Satya Nadella has openly suggested that SaaS as we know it might “collapse” in the agentic AI era.
So which future are we heading toward?
A world where anyone can build their own “ServiceNow at home”?
Or one where the big platforms become even more central?
Is This the Beginning of the End for SaaS?
Last year, Satya Nadella said something on a podcast that immediately sparked debate across the software industry. He suggested that in the agentic AI era, “the very notion that business applications exist” could eventually collapse. His reasoning was blunt: most business apps are still CRUD with rules. They store data and apply logic on top. What he believes will change is not the data itself, but where that logic lives.
Instead of being embedded inside every SaaS product, Nadella imagines that logic moving into an AI layer that operates across systems. Agents coordinating workflows, updating multiple databases, and executing multi-step processes end to end. In that world, apps become execution surfaces. The AI becomes the organizing layer.
ServiceNow’s Calm in the Middle of the Storm
This year, in a Business Insider interview, Bill McDermott (ServiceNow’s CEO) was asked whether customers could start building their own enterprise software with AI instead of buying platforms like ServiceNow. He said he is not worried… In his words, customers simply cannot recreate what took ServiceNow 20 years to build.
McDermott is not denying AI’s impact. In fact, he openly says AI is reorienting the global economy and enterprise technology itself. But his message is that there is a massive difference between using AI to enhance operations and replacing deeply embedded enterprise platforms that run at the core of the business.
The real inflection point: why build vs. buy is finally shifting
For decades, the decision to buy SaaS instead of building software internally was mostly an economic one. Building was expensive. Skilled developers were scarce. Internal IT teams were already overloaded. Buying software was often a faster, cheaper, and safer option for IT leaders.
Now AI is changing that equation. Even though it still takes senior engineers to reach true enterprise-grade standards, AI is clearly lowering the barrier to producing software in-house.
These new AI Tools are enabling a new type of builder. Not traditional engineers, but people from business, operations, and analytics who can now generate functional applications using natural language prompts.
These are what some call “vibe coders”, “AI-native developers” or “software composers.” They do not need years of computer science training. They need weeks or months of guided practice. That alone changes the supply side of software development in a fundamental way.
If a lot of people inside companies can suddenly build usable internal tools, the capacity to create software expands dramatically. What a single internal team can ship in a year will not look the same as it did just a few years ago.
This is where the build-versus-buy equation truly starts to wobble. Not because SaaS suddenly disappears, but because many projects that were never worth buying for now become cheap enough to build and experiment with.
At the same time, this new wave of AI-built software introduces a different kind of risk. Because while generating code is becoming easier, keeping that code secure, scalable, maintainable, and aligned with real business architecture is not. Without experienced developers in the loop, AI-generated apps can quickly turn into fragile systems full of hidden technical debt, security gaps, and performance problems that only surface later, when the business already depends on them.
Why “ServiceNow at Home” Is Still a Long Way Off
So, with everything AI now makes possible, is it realistic to build your own “ServiceNow at home”?
In theory, yes. In practice, it is highly unlikely. Not impossible, but extremely expensive, slow, and demanding in terms of long-term commitment. It would take many years, sustained investment, and a level of organizational discipline that very few enterprises actually have. So while it is not something we can fully rule out in the long term, it is clearly not a near-term reality.
For a very small slice of the market, it might happen. Large banks, big tech companies, hyperscalers. Organizations with hundreds of platform engineers, stable budgets, and a strong “build, not buy” culture. Many of them already run internal systems that resemble parts of ServiceNow: their own CMDBs, workflow engines, catalogs, and orchestration layers. AI can make that easier by speeding up UI generation, integrations, business rules, and documentation.
But for most companies, the real challenge is not building the first version. The real challenge starts the day after it goes live.
Because rebuilding something like ServiceNow is not just a development problem. It is a permanent vendor-level responsibility. Governance. Security. Compliance. Audits. Identity. Access control. 24/7 availability. Upgrades. Backward compatibility. Constant re-integration as the organization changes. That is the part AI does not magically erase.
When you decide to build your own platform, what you are really saying is:
“We are now the vendor. For years.”
For an average enterprise, that is a massive long-term burden. And it is precisely what ServiceNow abstracts away.
Conclusion: What AI Will Replace, and What It Won’t
This does not mean AI is about to replace software as a whole. A more accurate way to look at it is this: AI is far more likely to start replacing certain types of apps, not entire enterprise platforms.
The tools most exposed are narrow point solutions. Apps built to solve a single pain point. Internal dashboards. Simple workflow automations. Lightweight CRMs. Reporting tools. Small internal utilities created to unblock a specific team. These are exactly the kinds of systems organizations will increasingly build in-house with AI. Faster. Cheaper. And often “good enough” for their specific needs.
But confusing that with the ability to replace a platform like ServiceNow is a mistake.
ServiceNow is not just an app. It represents decades of accumulated learning about enterprise processes and workflows, built on top of massive volumes of real operational data. It is a platform configured by thousands of certified specialists across ITSM, HR, SecOps, CMDB, workflows, and automation. It is supported by vendor-grade security, compliance, upgrades, and a global ecosystem that keeps the platform running and evolving.
So yes. AI will reshape what companies choose to build versus buy. But replacing a platform that sits at the bone of enterprise operations is a very different challenge altogether…
And if scaling ServiceNow is among your priorities for 2026, as a ServiceNow partner, we are here to help you get the most out of the platform. Let’s talk.
