In a world where pretty much every app and service has massively changed over the years, browsers are something that have sort of stayed stagnant. You open them, type in a URL, and you get a search result. That’s it. There’s not much excitement or innovation left there. Or at least, that’s what I thought for the longest time, until AI browsers started popping up.
OpenAI was among the first companies to start hinting toward an AI browser, and while the AI giant only ended up releasing its browser, Atlas, a few days ago, a bunch of companies had already beaten it to the punch. For instance, Opera announced Neon, its futuristic AI browser, toward the end of May. Shortly after, Perplexity launched Comet, another AI-powere…
In a world where pretty much every app and service has massively changed over the years, browsers are something that have sort of stayed stagnant. You open them, type in a URL, and you get a search result. That’s it. There’s not much excitement or innovation left there. Or at least, that’s what I thought for the longest time, until AI browsers started popping up.
OpenAI was among the first companies to start hinting toward an AI browser, and while the AI giant only ended up releasing its browser, Atlas, a few days ago, a bunch of companies had already beaten it to the punch. For instance, Opera announced Neon, its futuristic AI browser, toward the end of May. Shortly after, Perplexity launched Comet, another AI-powered web browser. Since then, pretty much every other company seems to be racing to launch a browser built with AI at its core.
I’ve been testing all these browsers as they’ve rolled out, and Neon is currently my go-to. Although I’ve used these AI browsers for all sorts of tasks, my favorite use case so far has been using them to watch YouTube. However, while they’re great for it, that’s not where they truly shine.
Agentic AI is where things get really interesting
It actually does the work for you
Most of the AI browsers that have launched so far ship with something called Agentic AI, which is pretty much the next big step in the AI world. Agentic AI is essentially when artificial intelligence operates independently and actually executes tasks with minimal human intervention. To do so, it uses AI agents — systems capable of understanding context, making decisions, and performing actions on your behalf. From a user’s perspective, all one really needs to do is drop a prompt in completely natural language, and the agent can then come up with a plan to achieve whatever goal is set.
For example, you can tell an AI agent your vacation budget and ask it to plan an entire trip for you. You can even go one step further and ask it to book the trip lile selecting flights, hotels, and activities on your behalf. Finding the cheapest flight is a relatively redundant and boring task, and it’s one that can absolutely be handled by an AI agent without breaking a sweat.
What’s interesting is that AI agents don’t mindlessly keep executing a goal once it’s set. Instead, when they run into an error, they first try to resolve it themselves, often by retracing their steps or adjusting their approach. If the agent still isn’t able to fix the issue, it briefly pauses the task and asks the user for clarification or if they can perform something on its behalf.
For instance, I asked an AI agent to add some music to a Spotify playlist for me, and for some reason, the Add button wasn’t responding. Though the agent tried to get the button to work around fifteen times, it eventually ended up asking me if I could give it a look. Quickly refreshing the webpage did the trick! While concepts and early projects around Agentic AI started appearing months before users got their hands on AI browsers, they seemed to be primarily targeted at businesses and professional use cases back then.
With agentic browsers (which are AI browsers built around these autonomous AI systems), this technology has finally shifted toward everyday users like you and me. And honestly, the entire appeal of agentic capabilities is definitely best experienced within these browsers. It’s where you can actually see AI taking initiative, all in tasks you’d normally have to handle yourself.
You can automate pretty much anything with these browsers
The possibilities are endless
In theory, Agentic AI sounds cool. But then again, so does every other AI feature that gets announced. What’s great about Agentic AI, though, is that you can actually integrate it meaningfully into your daily workflow. For instance, being greeted by dozens of emails the moment you open your inbox is something everyone deals with, no matter what profession you’re in.
Even if you’re a student, your college inbox is always packed with random emails — trust me, I know. My college inbox currently has over a thousand unread messages. Unfortunately, most of the emails you receive are likely ones that don’t immediately need your attention. Another thing I do far too often is see an important email while having breakfast (when I’m quickly skimming through my inbox) and decide it’s probably better to reply when I’m at my desk. Of course, by the time I actually sit down, I’ve completely forgotten about it.
What I’ve done is set up an automation where my browser’s AI agent summarizes my Gmail inbox and highlights important unread emails from the past 24 hours. I’ve set this up as a scheduled task in Comet, and the AI browser automatically runs it every morning. It then notifies me with all the details I’ve requested in the form of a notification and also emails the summary to my inbox! Though this email automation is relatively basic, I also use the browser’s agentic AI capabilities to manage my very, very chaotic calendar.
For instance, manually time-blocking is something I spend far too much time doing — time that could be better spent on actually completing tasks, rather than just organizing them. With an AI agent, I can simply type out everything I’d like to time-block (or say it using the Voice mode), and it handles it all! I’ve even paired agentic browsers with my favorite AI tool, NotebookLM, which can intelligently analyze my calendar data and then suggest a realistic schedule based on when I work best. Together, they’re like having a supercharged personal assistant.
Another thing I use agentic AI a lot for is ordering coffee! I have the same coffee order, and I always order from the same place. So, why spend time placing the order when I can quickly have an AI agent do it for me? I have this saved as a shortcut too, and with just a single command, the AI agent places my usual coffee order automatically.
I can’t imagine working without it now
The tasks you can automate with agentic AI are quite literally endless. Of course, they vary from person to person, depending on their workflow. What I need automated might not be the same as what someone else does. The flexibility of agentic AI is exactly what makes it so powerful. The best way to figure out tasks you should automate is to glance at your current workflow, and then identify the repetitive, time-consuming, or low-priority tasks that eat up your day.