How AI Automation Disrupts Industries Built on Administrative Friction

The true impact of AI is not about making information easier to find. It is about making human labor easier to replace. By automating the administrative work that currently acts as a tax on the public, AI changes the game: it shifts the advantage from people who are good at managing complex systems to those who can simply define what they want to achieve. This reveals a simple truth: bureaucracy is often a gatekeeping tool that favors the educated middle class. When AI makes managing that complexity cheap, it does more than save time. It forces a change in how value is distributed. You should see this as a structural shift that will break or force a redesign of industries that rely on friction as a business model. The winners will be those who spot industries shielded by red tape and build tools that cut through it.

The Hidden Tax of Administrative Complexity

In their conversation, Tony Cowan-Bran and Benedict Evans point out that much of the modern economy relies on friction, inertia, and complexity. We often think this is because information is hard to find, but the real bottleneck is the human labor needed to navigate it. Whether you are trying to optimize a health insurance plan or manage credit card rewards, the system is set up to favor people who have the time and expertise to handle the paperwork.

> Bureaucracy is a regressive tax. To say bureaucracy privileges people with education, background, and experience is to say it is about managing complexity. It penalizes people whose experience is not in that field. Bureaucracy is a tax on the poor for the benefit of the middle class.
>
> -- Benedict Evans

This changes the goal from access to information to access to agency. The internet solved the library problem; AI solves the librarian problem. When you can delegate the task of filling out a 50-page form or comparing 20 insurance policies to an agent, you are not just saving time. You are bypassing the gatekeepers who profit from your inability to navigate the system.

Why Obvious Solutions Often Trigger Systemic Bloat

Many assume AI will make these processes leaner and more efficient. However, systems often react to efficiency by becoming more complex. When the cost of creating documents or arguments drops to zero, the volume of those documents explodes.

> You can make an argument that if AI makes complexity intelligible, the result is an explosion in complexity. And so you could have massively more granularity.
>
> -- Benedict Evans

We see this in legal sectors, where AI-generated filings are overwhelming courts. This creates a loop: as people use AI to create complexity, institutions must use AI to filter it. This creates a race where the baseline for what counts as a standard request keeps moving up. The advantage here does not go to those who use AI to do the work, but to those who build the infrastructure to handle the resulting scale.

The Analog Hole and the Limits of Automation

The most durable moats in this new environment are those that require a human signature or physical presence, known as the analog hole. Even as AI handles the research, strategy, and paperwork, the final gate remains the human who must legally authorize the action.

Evans notes that industries like banking often keep these barriers not because they are necessary, but as a way to restrict access. As AI gets better at mimicking human output, the value of the human in the loop will actually increase. The system will eventually route around automation by demanding higher-assurance, analog-verified authentication. The friction is not disappearing; it is just moving to the point of final authorization.

Key Action Items

• Audit your Friction-as-a-Service dependencies: Identify which of your business processes rely on customer inertia or administrative complexity. If a competitor can automate this away, your margin is at risk. (Immediate)
• Shift from Information Provider to Agent/Fixer: Move your product strategy away from providing data, which is now a commodity, toward providing the labor that executes the task. (Next 3-6 months)
• Prepare for the Granularity Explosion: If your industry involves high-volume documentation, expect an order-of-magnitude increase in incoming complexity. Invest in AI-native filtering systems now. (Next 6-12 months)
• Identify the Analog Moat: Determine where your industry requires human verification, such as signatures, notary, or physical presence. This is where the last, most stubborn friction will reside. (12-18 months)
• Prioritize Verification Over Generation: Build internal testing and human-in-the-loop verification layers into your AI workflows. The market will reward those who can guarantee accuracy, not just those who can produce output. (Ongoing)