The Shift from Seat-Based SaaS to Agentic AI
The traditional software-as-a-service (SaaS) model is facing a structural reckoning. For years, the industry relied on seat-based pricing—charging per user. Though, as AI shifts from a simple support tool to a primary driver of an “agentic world,” the value proposition of a user interface (UI) is diminishing.
Investors are increasingly concerned that agentic AI coding tools, such as Claude Code, will allow organizations to build their own custom software, effectively eroding the “moat” that software companies once enjoyed. If AI agents begin replacing human workers, the demand for individual software licenses will naturally plummet.
Why “UI Wrappers” are Failing and Data Hubs are Winning
Not all software companies are equally vulnerable. There is a stark divide between “UI wrappers”—simple interfaces that sit on top of other technology—and platforms that sit at the center of a customer’s data and workflow.
Organizations are generally reluctant to let AI agents run rogue or seize on the massive risk of software failure in their core systems. This creates a lasting advantage for vendors who maintain stability, safety, and security permissions.
The Quality Gap: The Hidden Risk of AI-Generated Code
While the bear case suggests that AI will replace professional software, the reality of code maintenance tells a different story. A study by Alibaba revealed that 75% of code created by AI models failed in less than a year, often because the models sacrificed long-term quality for immediate speed.
some early agentic tools have faced public criticism regarding security issues and reliability. This gap in quality ensures that professional, maintained software remains a necessity for enterprise-grade operations.
The New Orchestration Layer: How Palantir is Redefining the Model
While many SaaS stocks have struggled, Palantir has emerged as a signal for the future of enterprise AI. Rather than relying on a traditional SaaS seat model, Palantir functions as an “Operating System” for modern data and AI.
The company has moved away from slow, consultant-heavy pilots to a high-velocity Bootcamp strategy. This allows them to demonstrate real-world value using a customer’s live data in just five days, drastically shortening the sales cycle.
Real-World AI Orchestration in Action
Palantir utilizes four integrated platforms—Gotham, Foundry, Apollo, and AIP—to solve complex operational problems. Examples of this “Agentic AI” in practice include:
- Airbus: Managing complex global supply chains.
- NHS: Optimizing the availability of hospital beds.
- Wendy’s: Predicting inventory needs to reduce waste and improve efficiency.
The Strategic Moats of ServiceNow and Salesforce
Other legacy SaaS giants are pivoting to ensure they remain indispensable. The key is becoming the “master of records” for the data that AI agents need to function without hallucinating.
ServiceNow: The Technical Plumbing
ServiceNow’s strength lies in its configuration management database (CMDB), which serves as the heart of a customer’s technical infrastructure. By acting as the orchestration tool and plumbing for the entire software stack, it remains deeply embedded in business logic and audit trails. Its new AI Control Tower solution positions it as a major agentic AI orchestration platform.
Salesforce: The Data Engine
Salesforce has transformed into a master of records through its Data 360 solution and zero-copy technology, which allows it to capture data from cloud providers without expensive transfers. The acquisition of Informatica further enables the company to cleanse and govern data, providing the structured environment necessary for AI agents to operate reliably.
Frequently Asked Questions
What is the “SaaS-pocalypse”?
It refers to the market fear that agentic AI will destroy the traditional seat-based SaaS pricing model by allowing companies to build their own software or by reducing the number of human workers who need licenses.
Why is Palantir succeeding where other SaaS companies are failing?
Palantir focuses on the “orchestration layer,” integrating data and decisions rather than just providing a UI. Its Bootcamp strategy also allows for much faster value demonstration than traditional software pilots.
What is a “system of record” in the context of AI?
A system of record is a data source that is considered the authoritative version of a piece of information. AI agents require these clean, structured data sources to avoid hallucinations and ensure accuracy.
What do you think? Will AI agents eventually replace the need for third-party software, or will they simply make the “orchestration layer” more valuable? Let us know in the comments below or subscribe to our newsletter for more deep dives into enterprise AI.
