The SaaS landscape is experiencing seismic shifts. AI-powered interfaces are democratizing user experience design. Large language models are making sophisticated automation accessible to every company. Agentic systems are promising to revolutionize how we interact with software entirely.
In this new world, traditional competitive moats are eroding rapidly. The interface that took years to perfect? An AI can now generate a comparable one in minutes. The complex workflow automation that was your secret sauce? LLMs are making that trivial for any competitor to replicate.
But there’s one thing that remains unshakeable: the state.
The state is the anchor
When I talk about “state,” I’m referring to the core data model, the accumulated knowledge, the refined understanding of domain-specific problems that a company builds over time. This isn’t just raw data—it’s the structured, validated, and contextualized information that represents the real-world state of your customers’ businesses.
Consider a few examples:
Salesforce didn’t just build a pretty CRM interface. They built a comprehensive model of how sales processes work, what data points matter, and how to represent complex customer relationships. Their true moat isn’t the UI—it’s the decades of accumulated understanding about sales state management.
Stripe isn’t valuable because of their payment form UI. They’re valuable because they’ve modeled the incredibly complex state of global payment processing: fraud detection patterns, regulatory compliance across jurisdictions, and the intricate dance of money movement between parties.
HubSpot (where I work) didn’t win by having the best marketing automation interface. They won by building a sophisticated understanding of the customer journey state, how leads progress through funnels, and what data points actually predict conversion.
Why interfaces become commoditized
AI is making interface creation nearly frictionless. A well-prompted LLM can generate React components, design systems, and even entire application shells that would have taken teams months to build. The barrier to creating polished, functional interfaces is collapsing.
But creating an interface is the easy part. The hard part—the part that creates lasting value—is understanding what state that interface should represent and manipulate.
An AI can generate a beautiful dashboard, but it can’t generate years of domain expertise about which metrics actually matter for your specific use case. It can create elegant forms, but it can’t create the underlying data models that ensure those forms capture the right information in the right structure.
The state engine advantage
Companies that will thrive in this new landscape are those that have built sophisticated “state engines”—systems that don’t just store data, but actively model, validate, and evolve the real-world state of their domain.
These state engines have several characteristics:
Domain expertise encoded in data models: They understand the nuances of their problem space deeply enough to create data structures that reflect reality accurately.
Behavioral understanding: They know not just what data to capture, but how that data changes over time and what those changes mean.
Relationship mapping: They understand the complex interdependencies between different pieces of state and can maintain consistency across those relationships.
Validation and enrichment: They can automatically detect anomalies, fill in missing information, and maintain data quality at scale.
The irony of AI disruption
There’s an interesting irony here. The very technology that’s disrupting traditional SaaS models—AI—is also highlighting the importance of high-quality, well-structured state management. LLMs are only as good as the data they have access to, and agentic systems can only be as effective as the state they can read and manipulate.
Companies with sophisticated state engines often find their data becoming more valuable, not less, in an AI-driven world. They’re positioned to train better models, build more effective agents, and provide more accurate automation.
What this means for builders
If you’re building in the SaaS space today, ask yourself: What is the core state that my product manages? How deep is my understanding of that domain? How sophisticated are my data models?
The interface you’re building today might be obsoleted by AI tomorrow. But the state engine you’re building—the deep understanding of your domain encoded in data and behavior—that’s your lasting competitive advantage.
The winners in the age of AI won’t be the companies with the best chatbots or the most polished interfaces. They’ll be the companies that became the authoritative source of truth for their domain’s state. They’ll be the state engines that everything else connects to.
And in a world where everything can generate interfaces, being the source of truth is the ultimate moat.