AI Agents Are Already Changing Everything And Here’s How You Can Try Them Today

So you’ve probably been hearing a great deal recently about the amazing AI agents that will soon be transforming every aspect of our lives.

Whether we actually get real-life versions of sci-fi robots like C3P0 or Data, or just more useful versions of Alexa and Siri, AI agents promise to be a leap forward.

This technology will make it possible for machines and digital tools to carry out longer, more complex tasks, interacting autonomously with third parties on our behalf. This isn’t all just speculation—it’s happening now.

In fact, now is a great time to get to grips with it—and maybe even jump in and give it a go, if you want to make sure you’re ready when it really matters.

Agentic AI is already being adopted for commercial and industrial use cases. At this early stage, this usually involves deploying bespoke infrastructure, often in partnership with enterprise service providers like Amazon or IBM, and at a high cost.

The agentic tools currently available to consumers are, comparatively, at a very early stage of development. But even though they can be cranky, they offer a fascinating insight into how the way we use machines is changing, as well as an early chance to get hands-on with technology that many believe will define the next decade.

So let’s start with an overview of what’s out there right now, for those wanting to find out what all the fuss is about!

Agentic Options

There are two main instantly-accessible options for trying out agentic AI in a non-technical, ChatGPT-style way. These are the best options for non-techy types who just want to try it out.

OpenAI’s Operator is probably one of the most widely known. Be warned, it’s not cheap, though. Currently, it’s only available to Pro users paying the $200 monthly subscription in the US. It’s a web-browsing agent, meaning it uses computer vision to operate a browser and can carry out many of the same tasks we can. These include filling out online forms, logging into websites, booking flights and arranging for groceries to be delivered. Although slow at the moment, it’s an exciting preview of what we can expect from this tech in the near future.

Chinese challenger Butterfly Labs, which already upset the apple cart by developing its DeepSeek LLM for a fraction of the cost that was thought necessary, has also launched its own agentic option.

DeepSeek has already shown that it can build and operate LLM models at a fraction of what was believed to cost. It has also launched its own agent service, called Manus AI, at a more accessible price point of $39 per month. In demonstrations, Manus has been shown to be able to develop and deploy an entire website from a single prompt (if a little slowly).

At the time of writing, I was able to sign up for a Manus AI account for free by joining the waiting list here, but I have no idea how long this option will be available for.

Those not afraid of getting a little bit technical might want to check out AutoGPT. This is an open-source agentic framework that works with various LLMs to effectively give them agent capabilities. They do this by spinning up and coordinating the work of multiple genAI instances. It doesn’t have a friendly chatbot interface like ChatGPT, though, and needs to be built by pulling together models (GPT4, Claude etc), infrastructure (cloud or local servers) and data.

Those wanting to experiment with the potential for business transformation are likely to want to look at building their own agents to fit niche use cases.

The good news is that a growing number of solutions are appearing for those who might want to try this but don’t necessarily have the technical skills.

Microsoft AutoGen Studio offers a low-code interface for building and deploying agents that can perform business tasks such as finding new leads on LinkedIn or creating automated reports on the market and competitor landscape.

And Google’s Vertex AI Agent Builder builds on its Agent Development Kit framework to streamline the design, development and deployment of agents in businesses of any size.

It may be possible to put agents to work through your home assistant hardware and portable devices, too.

If you have a newer Amazon Alexa device, you might now have access to Alexa+, which adds some agentic capabilities, letting it understand longer commands, hold more natural conversations, and arrange for your oven to be repaired.

Hardware startup Rabbit’s portable genAI assistant was widely considered to be a disappointment, but the company is now focusing on developing agentic capabilities to make its devices more useful.

What To Watch Out For

AI agents are thought of as “next-level” AI because they represent a leap forward from the current generation of generative AI popularized by ChatGPT.

Adapting to keep up with the possibilities of this evolving toolset means staying up-to-date on the latest developments. Just as we saw ChatGPT evolve quickly to feature web search, image generation, speech, memory and vision, we can expect the usefulness of these new tools to improve rapidly, too.

But right now it’s early days. Many of these tools can be rough around the edges, they can (and frequently do) make mistakes, and according to most studies and metrics, are generally considerably worse than humans at most of the tasks they try.

There can be high costs, too. Pay-as-you-go services involving API use can quickly burn through tokens, leading to large bills.

And agents bring their own privacy issues to the table, too. As web browsing and computer-using agents can effectively have access to your entire screen, desktop or computer, special care should be taken to ensure you thoroughly understand what it’s using the data for.

Overall, even though the technology may be a little unstable and you probably don’t want to rely on it for anything critical, now is still a great time to start getting hands-on with agentic AI.

What we’re seeing today is just a glimpse of what will be possible in a few short years, and those with the best chance of benefiting from it will be those who jumped in at the start.