Imagine never having to relaunch an application.

Let’s say you are organising a vacation. Every small detail is taken care of, from the time you leave your house to when you board a train or aeroplane, from making hotel reservations to scheduling a taxi upon your arrival at your destination. Here’s the twist, though: you don’t have to launch any apps. Avoid switching between MakeMyTrip, Skyscanner, and Expedia. No downloading, no tab switching, and no juggling of five separate interfaces. Instead, all the heavy lifting is done for you by an AI agent that you are simply conversing with.

I realise it sounds insane. However, it is our destination.

What is an AI agent, then?

Consider them to be exceptionally intelligent digital assistants. When you assign them a task, they find a way to complete it. These agents are designed to work autonomously, whether that involves retrieving data from other apps, making quick judgments, or even acquiring new skills during their tasks. They do not require constant guidance in the form of instructions. Additionally, they remember their previous actions.

Assume that a business once required six employees to design and execute its marketing activities around the clock. However, an AI agent and one person can complete the task in less than an hour. Although that may sound like bad news for employment, there is a silver lining as well. These AI agents have the potential to revolutionise industries that are severely lacking skilled workers.

Consider cybersecurity. More than 5.5 million jobs in the field were unfilled worldwide in 2023. Businesses are exposed as a result. It’s difficult to identify and address threats in real time when there aren’t enough workers in the field. And things are just going to get worse. According to World Economic Forum research, the skills gap in cybersecurity has grown by an additional 8% since 2024. According to two out of every three organisations, they just lack the talent necessary to maintain their security. AI agents can help in this situation. They can indicate traffic surges when hackers attempt to flood a website with phoney visitors, prohibit hackers from accessing a system from strange places, and identify questionable logins. They may also move fast.

However, that only applies in the business sector. Artificial intelligence agents are as useful on a personal level. They can help you whether you need to create an image, evaluate a complicated spreadsheet, reset a password, or simply browse the internet. Additionally, you don’t need to use ten distinct apps to do a single task because they are adept at accessing other systems to retrieve the information they require.

With its “Buy for Me” AI shopping agent tool, Amazon is already experimenting with this concept. This AI agent will therefore search other websites for the product you’re looking for if Amazon doesn’t carry it. Additionally, it will place the order without you ever leaving the Amazon app if you like what it displays. For Amazon, that is fantastic. Not so much for the rest of us. Because other platforms may start to lose momentum if users stop switching between apps and only use the one where an AI agent handles everything.

And it is an issue. You see, apps are fuelled by user traffic. More users translate into more in-app purchases, sales, advertising money, and data for learning. Apps facilitate access to services. When using Adobe’s design tools, the app makes the process easier. However, the software loses business if users no longer feel compelled to access it.

Now, I am not bringing up this point merely because Amazon is developing a single AI agent. AI agents have been around for a while and aren’t particularly new. For comparison, it is anticipated that the global market for AI agents will reach a staggering $47 billion by 2030, up from $5 billion in 2024. That’s a 45% compound yearly growth rate!

There isn’t much discussion on how they’re changing customer behaviour, even if a large portion of this growth is on the corporate side. However, consumer-focused AI agents may gain traction quite rapidly if we were to learn from generative AI. For comparison, Gartner estimates that the emergence of generative AI and personal AI agents will result in a 25% decrease in search engine traffic by 2026. If that is any indication, you may find yourself using an AI agent instead of apps to complete tasks sooner than you anticipate.

So, will apps be discontinued?

You would think so. Particularly when you hear about AI agents like Midjourney that can assist you in creating beautiful visuals simply by conversing with them, you don’t need to become proficient in Canva or Adobe. Or Mindtrip, an AI travel planner that might take the place of your typical go-to websites like TripAdvisor or Expedia. Prominent figures like Nandan Nilekani and Satya Nadella predict that AI agents will eventually replace apps.

Here’s the problem, though. History demonstrates that technological advancements do not necessarily replace earlier developments. One thing history teaches us is that technological revolutions are rarely binary, as Senior Partner Chuck Whitten of Bain & Company so eloquently states. Usually, transitions don’t result in complete replacement. Rather, they establish ecosystems characterised by diversity—a blend of traditional and modern models, each occupying a specific niche.

Take a glance at the computer. Many people thought the personal computer would die when smartphones and tablets became so popular. However, the story you are currently reading was composed on a laptop. Computers did not go extinct. They recently discovered a new mission. When you have important work to do, they remain the preferred choice. They are designed for heavy-duty tasks that your phone isn’t capable of handling, such as multitasking and concentration.

Take Slack instead. It promised to transform communication in the workplace. People also believed that email would eventually become obsolete. But you know what? From 280 billion emails sent and received per day in 2018 to 360 billion in 2024 (a roughly 30% rise), email usage has only increased globally. By 2027, that figure is anticipated to reach 400 billion. Why? Because old habits don’t go away easily. Particularly in the business world, email still feels more formal and “official.” Furthermore, it is difficult to alter something once it has become the standard.

Perhaps AI bots and apps will cohabitate in this way as well. They might encourage apps to change rather than destroy them. To keep users in their ecosystems, we might see more apps creating AI agent-style capabilities of their own. Third-party artificial intelligence agents still require data, and the majority of that data is stored in apps that have been in use for years. These apps may provide richer, more integrated experiences than any external agent if they invest in developing their own intelligent AI interfaces.

And here’s something else. One AI agent might assist you with travel planning, but it won’t assist you with tax preparation or poster creation. Several AI bots would still be required for various tasks. That’s kind of the same as utilising different apps now. Indeed, the user interface may shift from tapping and swiping to just talking. However, the fragmentation is still present. There will always be trade-offs associated with convenience unless there is a single, all-powerful agent that can handle everything.

This is why apps may benefit from the entire AI agent wave. Indeed, to remain relevant, they may feel pressure to innovate, invest, and provide better experiences. However, that’s a good thing, isn’t it?