Artificial intelligence is no longer simply an enabling technology that slots neatly into existing enterprise software stacks. It is actively consuming them. What began with AI-driven copilots for writing, coding, or bookkeeping has evolved into full-stack replacements that directly threaten the very core of companies built on software subscriptions. The “AI eats software” thesis suggests that as models become more powerful and verticalized, the value of pre-packaged platforms—built on workflows, integrations, and incremental productivity gains—shrinks dramatically. Instead of buying a suite of tools, enterprises may increasingly rely on adaptive AI agents that assemble functions on demand. This shift directly puts companies like Adobe, Salesforce (CRM), Monday.com, Intuit, and others in the crosshairs.
Adobe once seemed untouchable, its Creative Cloud subscription model hailed as a fortress business. Yet, generative image and video models now allow a single prompt to accomplish tasks that previously required deep familiarity with Photoshop or Premiere. Already, startups like Runway, Stability, and MidJourney have shown that creative output can be streamlined to minutes, bypassing Adobe’s feature-dense interfaces. Unless Adobe re-anchors itself as the indispensable AI platform for creatives, its subscription model risks looking like legacy bloatware.
Salesforce, the king of customer relationship management, faces a similar reckoning. The very idea of a CRM as a static dashboard where sales teams log notes and track pipelines is antiquated when compared to AI agents that not only summarize calls, update records, and send follow-ups automatically, but also predict which leads are most likely to close. Instead of forcing companies to live inside a CRM, AI can act as the CRM—lightweight, invisible, and continuously adaptive. That undermines Salesforce’s lock-in strategy, where switching costs have historically been its greatest moat.
For Monday.com, the story is harsher. Project management platforms thrived by promising structured collaboration, but AI coordination agents are making them redundant. Why drag tasks across Kanban boards when AI can update project timelines, allocate resources, and notify teams in real time? The “collaboration software” category risks collapse into a thin layer of notifications on top of whatever AI engine a company chooses to deploy. The more powerful the coordination AI becomes, the more redundant a standalone project management platform looks.
Intuit, too, should worry. Its empire was built on digitizing complexity in tax filing, bookkeeping, and small business management. AI threatens to erase that complexity altogether. What happens when an AI agent can parse an entire year’s receipts, generate compliant tax filings, and offer real-time financial advice without needing TurboTax or QuickBooks? Intuit has brand recognition and trust, but as generative agents earn reliability through regulatory alignment, that trust advantage will erode. The company may soon find itself competing with AI-native tax advisors that cost a fraction of current subscription fees.
The broader reality is that software’s moat has always been friction: learning curves, entrenched workflows, and ecosystem lock-in. AI annihilates that friction by making advanced tasks instant, adaptive, and interface-less. The more AI learns, the less value is left for middleware software companies that package human workflows into SaaS models. The most resilient firms will not be those with the deepest existing customer bases but those that pivot fastest to owning the AI layer itself—turning software from a product into a continuously learning service. Everyone else will discover, too late, that AI doesn’t just integrate with software companies. It eats them.