Why The AI Era Is Unlike Any Technology Shift Before

The current AI era represents the latest and most rapid product cycle in technology, building upon the foundations of PCs, the internet, cloud computing, and mobile. Unlike previous waves, AI leverages the existing global infrastructure of cloud services and smartphones, leading to unprecedented speed in adoption and value creation. This shift is driven by a fundamental human desire to be “richer and lazier”—to generate more economic value while reducing manual effort. Enterprise spending on AI is inflecting sharply as companies move from experimental “magic tricks” to deploying solutions that save significant time and money.

Three Core Investment Themes in AI Applications

To build enduring, defensible companies in this new landscape, three primary investment theses have emerged:

1. Traditional Software Goes AI-Native

Just as cloud-native companies like Salesforce and Workday disrupted on-premise incumbents, a new generation of AI-native applications is emerging. The key opportunity lies not in “Brownfield” markets—trying to convince an existing customer to rip and replace a deeply embedded system—but in “Greenfield” markets. This includes:

• Net-new companies that adopt an AI-native solution from day one.
• Companies at an inflection point (e.g., outgrowing QuickBooks) that must choose a new, more powerful system.

For startups, the goal is to become the new system of record for a specific function (e.g., ERP, CRM, HRIS). These systems are incredibly sticky, creating high switching costs and turning customers into what can be described as “hostages,” ensuring long-term defensibility. While incumbents will certainly leverage AI to strengthen their offerings, startups have the advantage of building from a clean slate to win the Greenfield opportunities.

2. New Categories Emerge as Software “Eats” Labor

This is arguably the largest opportunity, as it involves competing not with the software market, but with the vastly larger labor market. AI makes it possible to build software that performs the job of a human, unlocking value in areas where the cost-to-value equation previously didn’t make sense for hiring a person.

• Case Study: Eve Eve builds a platform for plaintiff-side law firms, a market where attorneys operate on contingency and are thus highly motivated to increase efficiency. By owning the entire workflow from client intake to case outcome, Eve makes attorneys more productive, directly increasing their revenue. Its defensibility comes from creating a closed-loop data moat; the platform generates proprietary data on case outcomes, which is used to train its models to make better intake decisions—a compounding advantage that large, generic models cannot replicate.

• Case Study: Salient In the auto-loan servicing space, Salient provides an AI agent that replaces human call center workers. The core value proposition is not just cost reduction but superior performance, with the AI agent collecting 50% more revenue than its human counterparts. Its moat is built on two pillars:

  1. A data advantage: Having processed millions of calls, it knows the optimal script and approach for any situation.
  2. Product complexity: The system ingests and applies all federal, state, and county-level regulations in real-time, a task impossible for a human to manage consistently.

3. The “Walled Garden”: Applications Built on Proprietary Data

The third theme centers on companies that build defensibility by leveraging unique, proprietary data that foundational models do not have access to. AI transforms these businesses from mere data providers into finished product companies. Where a company like Pitchbook or FlightAware once sold raw data subscriptions, AI allows them to deliver a complete, high-value analysis, capturing a much larger portion of the value chain.

This proprietary data doesn’t have to be secret; often it’s public information that is simply difficult to aggregate, structure, and maintain over time. Examples include:

• Open Evidence: Built on exclusive licenses to the New England Journal of Medicine and other premier medical journals, providing a trusted AI tool for doctors.
• VLEX: Aggregated and digitized decades of Spanish legal records, then applied AI to quintuple its revenue by selling finished legal analysis instead of just access to documents.
• Ask Leo: Uses a proprietary trove of past procurement contracts to help companies negotiate better terms.

Building Defensible Moats in the AI Era

With the cost of building software plummeting, defensibility is more critical than ever. A crucial distinction exists between differentiation (a novel AI feature that can be easily copied) and defensibility. True, lasting moats are built on:

• Workflow Ownership: Becoming the indispensable system of record that a business runs on.
• Data Feedback Loops: Creating a closed system where using the product generates proprietary data that, in turn, makes the product smarter.
• High Switching Costs: Integrating so deeply into a customer’s operations that replacement becomes prohibitively difficult and costly.

• Tags:
• AI Applications
• Enterprise AI
• Defensibility
• A16z
• Investment Strategy
• Data Moats
• Systems of Record