Dan Faulkner is the CEO of SmartBear, helping teams build, test, and ship quality software.
getty
SaaS stocks have been under pressure as investors weigh the impact of an AI-driven "SaaSpocalypse." The theory is straightforward: If autonomous AI agents can create tickets, update records, reconcile invoices or resolve support requests without human intervention, then the SaaS interface—the very layer companies monetize—becomes obsolete.
From that perspective, enterprise software companies look like middlemen in a world where agents talk directly to each other.
That narrative is understandable. But it rests on assumptions that may very well turn out not to be true. Right now, at this early stage of the AI era, what markets are witnessing isn't the death of SaaS. It is uncertainty about how AI and agentic systems will reshape it. And when the future is unclear, investors retreat.
A core aspect of the "SaaSpocalypse" thesis is the assumption that the value of SaaS lies in its interface. It doesn't. The real value of enterprise software is in the workflows, domain expertise and structured data embedded inside those systems.
The assumption that AI can replicate 100% of SaaS features ignores the value of existing data, integrated workflows and—most importantly—trust.
Agents won't replace SaaS platforms; they'll interact with them differently. Most SaaS products have long exposed APIs that allow other systems to access their data and workflows. Agentic AI simply becomes another type of user, one that accesses the same capabilities through programmatic interfaces rather than graphical ones. Enterprise software will support two classes of users: humans and agents.
Another assumption driving the sell-off: AI will replace humans. This is the Luddite fallacy—the belief that technological innovation causes long-term net unemployment—rebranded for the AI age.
History tells a different story. Every time we've automated a core part of the software development lifecycle—moving from assembly to high-level languages, for instance—we didn't stop needing developers. We moved them farther up the stack to solve more complex problems. Despite headlines about AI killing developer jobs, recent research shows job postings growing.
AI is remarkable at creating code. But in doing so, it has also introduced more downstream problems: debugging, intent validation, faster and more complete testing to keep up with the faster code creation and architectural drift. These require more human oversight, not less. New categories of jobs and industry segments will emerge. Someone needs to address the security issues that agentic coding has created, for instance. These are opportunities for human ingenuity and employment.
Consider the efficiency paradox, what economists call Jevons Paradox. When people invested in making lighting more efficient to reduce energy consumption, the unintended consequence was that they used those efficient bulbs to light up everything, including spaces previously kept dark. As technology makes a resource more efficient to use, we don't use less of it—we use significantly more because it's cheaper.
If AI makes software cheaper to produce, demand will explode. We'll likely see a massive increase in the volume of software being built, creating more work across the software development lifecycle: security, privacy, end-to-end testing and observability.
The assumption behind much SaaS pessimism is that AI will spark a race to the bottom: companies delivering the same for less. In enterprise technology, the opposite dynamic usually wins. Vendors succeed by delivering "better for the same price" or even "better for more."
If agentic automation becomes widely available, it won't eliminate competition. It will raise the bar. Customers won't settle for cheaper mediocrity. They'll expect faster products, more reliable systems and better experiences.
There's also a misconception that once agents interact with other agents, execution becomes commoditized and the application layer loses value. The opposite may happen.
As autonomous systems proliferate, the most valuable capabilities will be governance, validation and control. Who audits agent behavior? Who enforces permissions and compliance? Who ensures that automated decisions remain transparent and accountable? These are problems SaaS vendors are well-positioned to solve.
Just because you can make your own bread doesn't mean you should. Testing applications, maintaining them, applying security updates, managing documentation—these are the unglamorous "cleaning the bread-maker" jobs of the software world. For most enterprises, paying someone else to handle them will be far more efficient.
SaaS vendors achieve economies of scale. They solve a problem once and amortize the cost across hundreds or thousands of customers. This makes buying more optimal than building at scale.
None of this means SaaS companies can ignore the shift underway. Those that fail to incorporate agentic workflows and modern AI into their products will struggle. Pricing models will evolve. I expect seat-based (subscription) pricing for people and consumption-based pricing for agents to coexist within the same product.
But SaaS players have the opportunity to tap two expanding budgets: traditional human-user software spending and a new category of agent-oriented IT investment. Rather than a zero-sum battle between AI and SaaS, we're likely to see rapidly expanding agentic workforces complementing a growing human one—raising expectations, accelerating innovation and pushing companies to solve problems better than ever.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。