The next internet won't be built for humans

For most of the internet's history, we've designed digital experiences around people. That sounds obvious, but it's worth reflecting on because it has shaped almost every decision we've made as technologists.

Websites were built so humans could find information. Search engines were built so humans could ask questions. E-commerce experiences were designed so humans could compare products, read reviews, make decisions, and complete purchases. Even when mobile arrived and fundamentally changed how we interacted with technology, the end user was still a person holding a device in their hand.

What changed over time was not who we were designing for, but the capabilities available to them.

The arrival of smartphones created an entirely new set of possibilities. Suddenly devices knew where we were. They had cameras, sensors, and constant connectivity. We could receive notifications wherever we happened to be. We could summon a car, transfer money, share photos, or pay for something with a tap.

Looking back, it's easy to see that mobile wasn't simply a smaller version of the desktop web. It was a fundamentally different environment that enabled entirely new types of businesses.

Uber couldn't have existed in its current form without location-aware smartphones. Instagram wasn't simply a website squeezed onto a smaller screen. Entire industries emerged because the technology created new capabilities that hadn't previously existed.

I think we're standing at a similar moment now with AI agents.

Most discussions about AI focus on the interface we can see. We talk about chatbots, assistants, copilots, and prompts. We debate which model is best and whether AI can replace particular tasks. Those conversations are interesting, but they sometimes distract us from a much bigger shift taking place underneath.

For the first time, we're beginning to create software that doesn't just provide information. We're creating software that can interpret information, make decisions, initiate transactions, and take action.

That distinction matters.

A human browsing a website wants an experience. They need navigation, visual hierarchy, trust signals, and content designed to help them understand what they're looking at.

An AI agent needs something different. It doesn't care whether a button is blue or green. It doesn't need a beautifully designed homepage. It doesn't need persuasive copywriting.

What it needs is structured information. It needs clearly defined actions. It needs trusted identity systems. It needs secure ways to access data and complete transactions.

In many ways, the internet we've spent decades building for humans isn't the ideal environment for software acting on behalf of humans.

That is why I believe we are about to see a significant shift in the infrastructure layer of the web.

The organisations laying the groundwork for this future aren't necessarily the startups making headlines. They are often the large technology and financial infrastructure companies that already sit at the centre of how the internet operates.

Payment providers are exploring how agents can securely initiate transactions. Identity providers are solving trust and authentication challenges. Cloud platforms are building environments where autonomous systems can operate safely. AI companies are creating the reasoning layer that enables agents to understand goals and determine actions.

Collectively, they are building the foundations of a web where software increasingly becomes a participant rather than simply a tool.

The implications for organisations are significant.

For years, digital transformation conversations have focused on improving human experiences. We have invested heavily in websites, mobile apps, customer journeys, and self-service capabilities. Those things remain important and will continue to matter.

What may change is the number of interactions that happen without a human ever opening a browser or downloading an app.

Imagine an AI agent that can compare insurance policies, book travel, manage subscriptions, negotiate utility providers, monitor spending patterns, and complete purchases on behalf of its owner. In that world, many decisions that currently happen through human interaction move into software-mediated interactions instead.

That doesn't eliminate the need for customer experience. It changes where customer experience happens.

Increasingly, organisations will need to think about whether their products and services are accessible not just to people, but also to the software acting on people's behalf.

Can an agent understand your products?

Can it access the information it needs?

Can it complete transactions securely?

Can it trust the systems it's interacting with?

These might sound like technical questions today, but I suspect they will become strategic business questions sooner than many organisations expect.

The mobile revolution created businesses that couldn't have existed in the desktop era. Not because the information wasn't available, but because the technology environment hadn't yet provided the necessary building blocks.

The same pattern may emerge again with AI agents.

The most successful companies of the next decade may not simply be using AI to improve existing processes. They may be building entirely new products, services, and business models that only make sense in a world where software can read, decide, pay, and act.

If that happens, the next internet won't stop serving humans.

It will simply have a new type of user alongside them.