What happens to mobile apps when AI stops opening them?

The mobile economy is built around a simple habit: open the app.

Need a ride? Open Uber.
Need groceries? Open Instacart.
Need to book a flight? Open an airline or travel app.

Every digital product was designed around that moment. The interface guides you through options, helps you compare choices, and eventually leads you to a decision. That structure works when humans are doing the deciding.

But a growing number of systems are starting to make those decisions themselves. And once that happens, the reason many apps exist in the first place begins to disappear.

The shift is already underway

You can already see the early signals across industries. In insurance, Lemonade uses AI systems that can review and approve certain claims in seconds, issuing payment without a human adjuster involved. In healthcare, tools like Abridge now listen to doctor–patient conversations and automatically generate clinical notes, referrals, and billing documentation.

Entertainment is evolving in a similar direction. Spotify’s AI DJ curates entire listening sessions based on a user’s taste and behavior, effectively deciding what plays next without requiring the listener to search or browse. Netflix’s recommendation systems now influence the vast majority of what viewers watch.

Even mobility works this way. When you open Uber, the most important decisions have already happened. Algorithms determine which driver is assigned, what route they take, and how the ride is priced before the rider interacts with the interface.

None of these systems look like the cinematic version of autonomous AI agents. But they share the same pattern: software is increasingly coordinating outcomes rather than simply presenting options.

Most apps exist to help people decide

A surprising number of mobile apps exist primarily to help humans make choices.

Travel apps help compare flights.
Retail apps help browse products.
Delivery apps help choose restaurants.

These interfaces exist because humans need help navigating information. But that is exactly the kind of work machines are built to handle.

An agent can monitor prices continuously, evaluate hundreds of options instantly, and execute decisions based on preferences a user has already defined. The outcome still happens, but the user no longer needs to navigate the interface to make it.

Once that shift takes hold, many apps start to look less like destinations and more like infrastructure.

Services become ingredients

When systems begin assembling outcomes across multiple providers, individual apps stop being the place where the experience lives.

A rideshare platform becomes a transportation provider that a system calls when needed. A retailer becomes a supplier inside an automated purchasing workflow. A financial service becomes infrastructure that a personal finance system interacts with directly.

Amazon has already started experimenting with this model, testing systems that can purchase products from other retailers on a customer’s behalf when the item is not available on Amazon. The service still matters. But the interface may not.

The real battle moves upstream

If this model spreads, companies are no longer competing primarily for attention inside their app. They are competing to be selected by the systems acting on behalf of their users. And that fundamentally changes how digital markets work.

Success becomes less about engagement metrics and more about reliability, consistency, fulfillment rates, pricing stability, and predictable outcomes. The services that perform best inside automated systems will increasingly be the ones those systems choose.

In other words, the most important customer interaction may no longer happen inside the app. It may happen inside the system deciding which service to use.

The next control point

This is why so many of the world’s largest technology companies are racing to build assistants that sit between people and the services they rely on. Whoever controls that orchestration layer effectively shapes the market.

We have seen versions of this before. Search engines determined which websites people visited. App stores determined which software people downloaded. Agentic systems may determine which services people use.

And if that happens, the biggest shift in the digital economy will not be better apps. It will be the realization that we need far fewer of them.