Palo Alto Networks CEO Nesh Aurora didn't mince words on the All-In Podcast: “If you're an analytical SAS company, it's over.” He sees a swift, AI-driven apocalypse coming for entire categories of enterprise software, fundamentally reshaping how founders build and profit.

Key Takeaways

  • Analytical SaaS is extinct: Companies whose core value is reporting and analyzing data face an existential threat because AI models can perform these functions directly, eliminating the need for expensive, incremental software modules.
  • UIs will vanish for 'systems of work': Traditional enterprise UIs, particularly for systems of work and record like HR or CRM, will be replaced as AI agents take over complex tasks, shifting user interaction from screens to agent directives.
  • Data infrastructure explodes: Enterprises will require ten times more data storage in the next three years, making databases and underlying infrastructure more essential than ever, despite potentially being undervalued.
  • New profit pools emerge in 'harnesses': The real opportunity shifts to specialized application layers that build "harnesses" for AI models, abstracting their complexity and integrating them into agentic, domain-specific solutions.

The SAS Apocalypse Arrives Early

Aurora's vision for analytical software is bleak. He argues that the entire premise of analytical SaaS—selling modules to crunch data and provide insights—is obsolete. Why? Because large language models (LLMs) can do it cheaper, faster, and more directly. “You don't need that. I can just go run NLM against the data,” Aurora explained. This means the hundreds of millions of dollars spent on enterprise dashboards, reporting tools, and incremental analytic features suddenly provide diminishing returns.

For founders building in this space, it's a stark warning: if your product's primary function is to help humans interpret data that an AI could directly query or summarize, your clock is ticking. The value isn't in the analytics; it's in the raw data itself, and the models that act on it.

UIs Go Away, Agents Take Over

Beyond analytics, Aurora predicts a dramatic change for systems of work and record. Think HR systems, CRM, ERP—the sprawling, often clunky enterprise software with complex user interfaces. His forecast: “If that happens UI goes away.” The idea is simple: why navigate a maze of menus and clicks when you can tell an AI agent what you want done?

“I just tell an agent look figure out from my sales call figure out the key points and go post it into you know whatever sales tracking system I have with this Oracle or Salesforce,” Aurora said. “An agent conceptually should be able to do it.” This shift means the lucrative business of building and maintaining these UIs will erode. The new battleground will be in creating the specialized applications that enable these agents, the "harnesses" that integrate AI capabilities into specific business functions. Companies won't write every piece of software; they'll want their HR system, agentic-enabled and AI-enabled, delivered by a specialized application vendor.

The Unsung Hero: Data Infrastructure

Despite the dramatic shifts in application layers, one area will see explosive, if undervalued, growth: data infrastructure. Aurora notes that for all this AI to work, it needs immense amounts of data. “We are going to need 10 times the data stored in enterprise than we have today for the next three years,” he stated. This means databases, storage, and data pipeline technologies become more critical than ever.

Founders in infrastructure should prepare for a surge in demand, but also understand that while essential, the pure "plumbing" might not command the same premium as the agentic application layer. The value will be in reliably capturing, storing, and serving up the raw material that fuels the AI revolution.

What to Do With This

Pull your product roadmap for the next 12-18 months. If your core value proposition relies on humans actively analyzing data or clicking through complex UIs to complete routine tasks, pivot now. Re-evaluate your product through the lens of an AI agent: can an agent do this work? If so, your opportunity lies in building the specific agentic "harness" or data infrastructure that makes that agent effective, not the analytical layer or the human-facing UI.