Does the rise of AI agents signal the end of the app economy?

The app economy is rapidly coming to an end.

It’s probably a sensational sentiment, but let’s play out this scenario anyway.

Over the past year, I’ve been closely tracking how much of the enterprise software and SaaS services we use are being deprecated into data services and databases. AI tools and agents remove layers of abstraction and enable single-screen type use cases where multimodal generation tools can be leveraged to build real-time UIs for specific queries or use cases. (Think search with Gen AI for business.)

Now shift your thinking to the app experience most of us have today when interacting with a mobile device or tablet. Many of us use dozens, if not hundreds, of applications for everything from banking to travel to work to gaming.

To make everything work, I organize my apps into folders and use Face ID for easy login access. In the process, we give these apps pure access to our data to ensure the experience is good, but often with little consideration for how the data is used (see this (A topic for another time). App builders have built apps to improve the UI/UX and connect to data services such as rideshare, hotel and flight, food, and investment services.

So imagine the AI-powered devices of the future, with powerful GPU/NPU-equipped security operations centers and greater connectivity that allows digital assistants and agents to operate at the operating system level. Just talk to your digital assistant.

In my opinion, this is where there is a huge opportunity for Apple intelligence, allowing Siri (or whatever it will eventually be called) to understand your requests, sift through your apps, and provide you with the best experience. Cross-reference things. A similar “super agent” will exist for Android as well.

The ultimate journey (app)

Currently, if you want to book a trip, you can use integrated services like Google Flights or apps like Expedia. These apps have access to extensive datasets and user interfaces that make it easy to book flights and travel. As users, we are limited to that app’s data services and the group of apps we can use to compare with each other (like Google Flights or Lending Tree).

In the future, we will be able to use multimodal (text, voice), OS-level digital assistants depending on our needs. Imagine being able to do that. For example, you can book a flight to San Francisco, a hotel until Thursday, a driver to pick you up when you arrive, and reservations for dinner on Wednesday. It has all these features. To confirm meetings and other tasks.

Assistant + Agent technology will help you navigate based on your preferences, including which airline you want to fly on based on your status, travel budget, seat preferences and availability. Complete tasks with secure payments. Choose your favorite ride-sharing or ride-hailing app, or choose your preferred hotel based on room type, amenities, services, and location based on your meeting.

The same principle applies to restaurant reservations, and of course it’s easier to confirm meetings, but it always helps to make sure you’re traveling for a legitimate reason. Plus, you can tap into and search for data and services you don’t already have an app for, find better deals on shopping and travel, and suggest better dinner places based on the preferences of the people you’re meeting. You can also. .

Why do we need an app? At least that’s what we know. Will apps retire data services and user interfaces that can be generated in real-time?Will apps be retired, or will they become more of a data layer, as business apps will become?My preference is for apps Not. It’s the experience that the app creates. Call a car, prepare some coffee, and show me your account balance.

Enterprise apps: which one to deploy first?

And we’ve all heard at some point that the enterprise app layer is changing. Microsoft, Salesforce, ServiceNow, Oracle, and SAP, to name a few, are all fighting to become a single pane of glass to help customers be more productive and efficient. Copilot and Assistant are extended to multi-threaded inference agents that can navigate and arbitrate across a large amount of data services that will likely soon be obsolete.

We know that business applications often suffer from outdated user experiences that are difficult to navigate. Instead of moving between dozens of applications such as CRM, ERP, HCM, SCM, etc. in this area, the near real-time AI-generated UX is continuously adapted and optimized based on the user’s preferences.

UI/UX abstracts the connected open data ecosystem and allows its needs to be clearly expressed. However, a multimodal future is not far away, where agents can be entrusted with tasks simply by telling them what they need.

Apps will become a thing of the past. The only question is when.