Key Takeaways

  • Palo Alto Networks CEO Nesh Aurora predicts a "SAS apocalypse" for analytical software, stating that traditional UI-driven enterprise tools are effectively "over" because AI models can perform data analysis directly.
  • Aurora argues that “UI enterprise software and consumer software UI is the worst thing we did as technologist,” advocating for a shift to AI agents interacting directly with data rather than through complex human interfaces.
  • Jason Calacanis provided a tangible example of this shift, reporting a 90% reduction in a company's SAS bill by replacing 17 accounts with three AI-connected ones, managed through natural language via Slack and Claude.
  • The next five years will see a radical "re-engineering" of "system of work" SAS, as AI agents aim to achieve true efficiency by reducing human effort (e.g., "five people become one in a company").
  • Businesses relying on analytical SAS for data interpretation will find their value eroded as customers move to AI models that handle both data collection and analysis internally.

The Method: How AI Agents Are Killing Analytical SAS

Nesh Aurora, CEO of Palo Alto Networks, doesn't mince words: “If you're an analytical SAS company, it's over.” His stark prediction stems from a clear observation: AI models are becoming so capable that they can collect and analyze data directly, rendering many intermediary software modules obsolete. Aurora explained his perspective plainly: “I'm going to collect a lot of data for you and analyze it for you. I don't need you to analyze it for me. I can run models against data and analyze them myself.” This isn't just theory; it's a strategic shift happening right now. For years, businesses bought SAS tools to visualize and make sense of their data. Now, the AI itself becomes the analyst.

He sees traditional UI-driven enterprise software as a major impediment to efficiency, calling it “the worst thing we did as technologist.” The future, according to Aurora, involves AI agents interacting directly with systems and data, bypassing the need for human-centric interfaces. This isn't just about analytics; it's about a complete reinvention of "system of work" SAS over the next five years. The goal is to achieve "true efficiency where five people become one in a company," a level of productivity only possible when AI streamlines workflows at a fundamental level.

Jason Calacanis provided a real-world illustration of this transformation. His team reduced a substantial SAS bill by a staggering 90%. They achieved this by “created like three accounts, got rid of 17, connected it to Slack, connected it to Claude, and now everybody can interface it through a natural language.” This shows how AI agents, acting on natural language commands, can consolidate and automate tasks previously requiring multiple distinct software subscriptions and human interaction with their UIs. The immediate impact on cost and efficiency is undeniable.

Where This Breaks Down

Aurora's vision holds immense weight for specific types of software, particularly those focused on data analysis and basic workflow automation. However, this "apocalypse" isn't a blanket condemnation for all SAS. Highly specialized software with complex visual components, such as CAD programs for engineering, video editing suites, or intricate design tools, will likely retain their UI-driven nature. For these, the visual interaction and human intuition guided by a sophisticated interface remain essential. Imagine trying to design a circuit board or edit a film through natural language prompts alone; the nuance and precision a UI offers would be impossible to replicate efficiently.

Furthermore, legacy systems and heavily regulated industries face significant hurdles in adopting full AI agent-driven workflows. Data silos, compliance requirements, and the sheer cost of re-engineering entrenched infrastructure mean the transition will be slower and more complex. The immediate 90% cost savings Calacanis saw might not be easily replicated where data access is fragmented, security protocols are rigid, or human oversight is a legal necessity. There's also the challenge of 'explainability' – if an AI agent makes a decision, understanding why it did so can be critical for auditing or debugging, and current AI systems aren't always transparent enough.

What to Do With This

If you're a founder building an analytical SAS product, immediately re-evaluate your core value proposition. Can your customers achieve the same insights by directly feeding their data into an LLM or specialized AI agent? If so, you're competing against a nearly free, infinitely scalable analysis engine. Shift your focus to data acquisition, data quality, or actionable insights derived from AI outputs, rather than providing the analytical engine itself.

For any founder using a significant amount of SAS, conduct a "Calacanis audit" this week. List your top five most expensive analytical or workflow automation SAS tools. Explore whether off-the-shelf AI agents (like those integrated with Slack or Claude) can replace or consolidate multiple subscriptions. Start with simple tasks – report generation, data extraction, or basic process approvals – to identify immediate cost-saving and efficiency gains. The goal isn't to eliminate UI entirely tomorrow, but to strategically prune your software spend where AI agents already outperform human-UI interactions.