The term "system of record" is an outdated metaphor, like the floppy disk save icon. Atlassian's CEO argues that modern knowledge businesses are dynamic collections of processes. The true value lies in coordinating these processes efficiently, not just storing data. AI's role is to orchestrate this flow.
While replacing complex systems like Workday with AI is impractical, the real opportunity is in extensibility. AI allows users to build small, custom apps on top of existing platforms, solving specific needs and making the core SaaS product even stickier and more valuable.
Drawing on Dan Ariely's "Predictably Irrational," per-seat pricing succeeded because it feels psychologically fair. Customers are more willing to pay for perceived effort or scale (more employees = more cost) than for brutally efficient outcomes, as illustrated by the locksmith paradox.
For 60 years, software digitized physical filing cabinets into databases, improving data retrieval but not the work itself. The current AI wave represents a paradigm shift where software (the "filing cabinet") can now autonomously perform tasks, evolving from a system of record to a system of action.
SaaS companies are not equally vulnerable to AI. Some (like Zendesk) tie seats to work AI can replace. Others (like Workday) use seats as a proxy for company size and are safer. Markets are currently failing to differentiate, creating a valuation gap worth understanding.
The idea that companies will use AI to build their own enterprise software is flawed. It ignores the vast number of non-obvious edge cases (e.g., state-specific labor laws) that mature SaaS products have codified over years. This accumulated, deterministic logic is a powerful, hard-to-replicate moat.
An overlooked benefit of per-seat pricing (e.g., Workday) is predictability for the vendor's sales team. Sales leaders can accurately forecast deal sizes based on a prospect's public employee count, making it far easier to scale a sales organization efficiently compared to unpredictable consumption models.
Usage-based pricing for AI faces strong customer resistance. Unlike cloud storage where usage is directly controlled, AI credit consumption can be driven by new vendor-pushed features. This lack of control and predictability leads to bill shock, making customers prefer the stability of per-seat models.
AI model capabilities have outpaced their value delivery due to a fundamental design problem. Users are inherently scared and distrustful of autonomous agents. The key challenge is creating interaction patterns that build trust by providing the right level of oversight and feedback without being annoying—a problem of design, not technology.
Mike Cannon-Brookes suggests viewing business functions through a new lens. "Input-constrained" work (e.g., customer support, legal) has a finite queue; AI drives cost efficiency. "Output-constrained" work (e.g., R&D, marketing) is limited by creativity; AI can be reinvested to generate more value.