For years, we have echoed a phrase that resonated deeply: “software was eating the world.” This simple assertion encapsulated how nearly every sector became intertwined with applications, platforms, and cloud services. However, a seismic shift is occurring, one that is just as profound: the artificial intelligence (AI) revolution is not merely transforming entire industries; it is pressuring the software industry from within. This brings forth a provocative question: if AI can now craft custom tools in mere moments, why continue investing in rigid, standardized software that often constrains how businesses operate?
At this juncture, the debate intensifies. We are not merely discussing incremental improvements; we are reevaluating the current model for enterprise software. The possibilities for dramatic transformation are on the horizon. The key word here is “potentially.” There exist compelling reasons to believe in this change—and equally strong arguments for its inherent limitations.
Software in the Age of Artificial Intelligence
Consider a fundamental question: what are you truly paying for when you invest in software? Traditionally, the cost incorporated the development of the tool, ongoing evolution, and the labor required to render it sufficiently generic for a wide audience. With AI’s capacity to generate code quickly and economically, the perceived value begins to shift. It moves towards flow design, seamless integration with existing business systems, and deliverable, measurable results. Bret Taylor, founder and CEO of Sierra and a board member of OpenAI, emphasizes that the focus should be on the value delivered to the client, not on technology for technology’s sake.
Previously, businesses faced a straightforward choice: purchase a pre-packaged tool and accept its limitations or invest in a custom development that often proved costly and slow. What AI introduces is an alternative approach. Instead of selecting software, companies might simply articulate their issues, allowing AI to design, deploy, and fine-tune custom systems as processes evolve. As Bret Taylor notes, “Our hypothesis is that, if we progress five years into the future, most digital interactions will be managed through an agent.” If accurate, this could fundamentally redefine the dominant interfaces companies use today.
Significantly, this dialogue is no longer confined to industry conferences or investor chats. Substantial shifts are becoming apparent. The phenomenon known as “vibe coding” has emerged, allowing even non-developers to create websites or applications simply by describing their needs verbally. Platforms such as the European Lovable have championed this concept, resulting in fewer technical barriers, rapid iterations, and a culture of experimentation rather than lengthy projects. However, this does not imply that businesses will replace their ERP systems with instant, AI-generated solutions. It does illustrate why the market is contemplating these possibilities seriously.
Yet, enthusiasm often meets the harsh reality of enterprise needs. Corporate software is not standalone; it is deeply integrated with databases, legacy systems, user identities, permissions, compliance checks, and audit trails that have evolved over many years. Moreover, the critical aspects of regulatory compliance and security pose real challenges. Even if an AI-designed system appears functional, essential questions remain: Who will maintain it? What happens when it breaks? Who will ensure it remains operational over time? Although the allure of “customized and rapid” software is compelling, significant challenges still lie ahead.
If this conversation feels abstract, Bloomberg offers a tangible insight: the market is starting to react as if these threats are real, though its ultimate impact remains unknown. The report indicates that Anthropic’s release of Claude Cowork has reignited fears of disruption posing a challenge to traditional software. According to analyses by Morgan Stanley, indices tracking SaaS have decreased by 15% in 2026, following an 11% drop in 2025—the most severe decline since 2022. Some analysts even assert that there are currently no compelling reasons to invest in software companies.
Images | Hack Capital | Anthropic
In Xataka | Meta was the big loser of the AI race in 2025. She was actually preparing her big move.
In Xataka | AI has already disrupted the world of programmers as we knew it. Now the challenge extends to translators.