Is 2025 Really the Year of Agentic AI?

In January, I wrote that AI agents were about to be the next big thing for artificial intelligence.

That article included a prediction from Nvidia’s CEO Jensen Huang who said: “I think this year we’re going to see it take off.”

He also called agentic AI a “multi-trillion dollar industry” in the making.

Now that we’re halfway through the year, let’s check-in to see if Huang and I are right.

Are AI agents really taking off?

How far have we come in just six months… and what still hasn’t quite materialized?

A Mixed Bag With a Promising Future

When we talk about “agentic AI,” we’re talking about AI that doesn’t just answer questions…

It also takes actions on your behalf.

For example, it can book a flight for you, manage your calendar, send emails, build apps or automate multi-step tasks based on simple instructions.

At least, that’s the vision for agentic AI.

But this vision is not quite a reality yet for everyday Americans. At least, not at scale.

I’d bet that most people haven’t interacted with a true AI agent yet.

They’ve used chatbots, even some incredibly smart ones. But most people haven’t experienced AI tools that can proactively handle multi-step tasks.

Readers of this letter might recall the Rabbit R1, a $199 pocket-sized AI device that promised to replace your phone’s apps using a “large action model.”

It shipped in April 2024…

And I sent mine back in a week.

The R1 could only perform basic tasks like setting a timer or playing music. It relied on scripted demos. Reviewers called it “a glorified button.”

And it flopped. But the idea was solid.

I believe we will all soon interact with a physical agent that acts like a personal assistant. But we aren’t there quite yet, either technically or commercially.

That said, Google Gemini Live is rolling out this summer.

It promises an on-phone conversational agent with memory, voice interaction and real-time reasoning. Early demos show it planning trips, pulling up documents and even adjusting its tone based on your mood.

But it’s not fully live for most users. And while the demos look cool, it looks like this AI still needs you to drive the conversation. Meaning, it won’t plan your day unless you ask…

Which makes it more of a smart assistant than a true AI agent.

But Anthropic’s Claude AI recently launched something closer to a true agent, called Workflows.

These are customizable, multi-step automations that can monitor documents, trigger actions and call third-party tools.

But setting them up still requires a bit of technical know-how, which puts them out of reach for most people right now.

Not to be outdone, OpenAI acquired Jony Ive’s startup io Products for $6.5 billion in May,

Ive was the chief designer of the early iPhones. Now he’s developing devices tailored for AI applications.

I’m curious to see what he and the OpenAI team will come up with. In the meantime, we’re still waiting for a true AI agent that lives on our phones, understands our preferences and gets things done without a lot of hand-holding.

But while we wait for these commercial applications, we’re seeing major progress in more technical areas…

Especially in software development.

On a Reddit thread asking “What is one AI Agent you cannot live without?,” many responses from software developers were similar to this one:

Being able to increase your productivity 10X seems like a dream come true.

But soon AI agents might replace coders entirely.

One of the most impressive examples of an AI agent released this year is Devin, created by Cognition Labs.

Devin functions as a full software engineer. It can read a bug report, search for solutions online, write code, run it, debug it and then deploy a working version.

And it does all of this independently.

What’s more, it shows its reasoning step-by-step as it works.

It isn’t perfect, but it’s a huge leap forward from the autocomplete-style tools that most developers use today.

Not surprisingly, Devin has also started attracting some serious attention from major tech firms.

Then there are tools like Replit’s Ghostwriter, Zapier’s AI Actions and Microsoft’s Copilot Studio that are layering agentic capabilities into existing platforms.

These tools enable users to chain actions and automate tasks, making them much more efficient at their jobs.

Even Salesforce has begun testing “digital employees.” These are agent-style tools that monitor workflows and step in automatically when something breaks.

But these tools are mostly limited to developers and IT teams.

They aren’t impacting the average person’s daily life yet. Although, they offer a clear picture of what’s coming…

And very soon the rest of us could have access to AI agents that are just as impactful.

Especially after what happened this past week.

Last Thursday, OpenAI announced a new feature inside ChatGPT called Agent.

It was quietly released to a handful of early users…

But it could turn out to be the most significant feature OpenAI has launched to date.

You see, ChatGPT Agent is designed to act like a real assistant.

It’s not just something you talk to…

It’s something that takes action for you.

You can set one up to plan a trip, organize receipts or even launch parts of a marketing campaign.

And you can do it without writing any code.

Right now, ChatGPT Agent is only available to a limited group of enterprise users.

And as this headline from Futurism.com indicates…

Results so far have been a mixed bag.

But it could soon become the closest thing yet to the vision Jensen Huang and I described six months ago.

OpenAI’s slow, deliberate rollout of this new feature suggests that the company knows Agent is something big…

The first step toward a new kind of computing experience.

Here’s My Take

Back in January, I said 2025 would be the breakout year for AI agents.

For most people, we’re not quite there yet.

But progress is moving faster than many realize.

And while the first half of the year has been quiet, the second half could bring an explosion.

Developer tools are reshaping the software industry. New consumer tools are coming.

And OpenAI’s new Agent feature may finally open the door to true AI-powered assistants.

An always-on, personalized AI agent that lives on your phone, knows your needs and gets things done even while you sleep.

And when that happens, Jensen Huang’s “multi-trillion dollar” prediction won’t sound so bold after all.

Regards,

Ian KingChief Strategist, Banyan Hill Publishing

Editor’s Note: We’d love to hear from you!

If you want to share your thoughts or suggestions about the Daily Disruptor, or if there are any specific topics you’d like us to cover, just send an email to [email protected].

Don’t worry, we won’t reveal your full name in the event we publish a response. So feel free to comment away!