TMT Predictions 2025: Insights from Deloitte
06:00 Share

Okay, got to admit, I've been kind of glued to all the AI news lately, especially anything about AI agents. You know, there's a ton of those programs that can handle tasks and make decisions, even chat with us. Like, they're right out of science fiction. But it's easy to get lost in all the hype, right? So in this deep dive, we're going to get practical. We're going to look at what's really happening with AI agent adoption in businesses beyond headlines. And to kick things off,

We've got some fascinating insights from NVIDIA. You know they're known for their graphics cards, sure. But they're making a serious push into the AI world, especially with these things called microservices for AI agents. Now stick with me here because this is where it gets interesting.

These microservices are like tiny software building blocks, and they're designed to make AI agents safer and more reliable for businesses. One of them tackles the problem of harmful or biased outputs. Think of it as like a content safety net for AI conversations. Another microservice keeps AI agents focused on specific topics, making sure they don't suddenly veer off into like irrelevant tensions. And here's the one that really caught my eye. A microservice designed to prevent jailbreaking.

Jailbreaking in this context means someone trying to remove the limitations on what an AI agent can do. It's like trying to hack your phone to bypass the parental controls, but with potentially much bigger consequences. NVIDIA is clearly thinking ahead here, anticipating the risks and trying to build in safeguards.

But here's the thing. NVIDIA isn't just quietly rolling out these safety features. They're making a big deal about it. And that tells me something important. They're betting that safety and control are going to be the key factors that convince businesses to really embrace AI agents. It's a clever strategy, really. Because, let's face it, if you're a company and you're considering handing over tasks to an AI, you want to know it's not going to go rogue and start spouting conspiracy theories to your customers or make decisions that could land you in hot water.

But hold on. Before we get too carried away with NVIDIA's vision, we need to look at the bigger picture. Are businesses actually adopting AI agents as quickly as all the hype suggests? Well, that's where things get a little complicated. We've got some conflicting predictions here. On one hand, you have Salesforce's CEO, Mark Benioff, basically saying that AI agents are about to explode. He breaks a billion of them running on their platform alone in the next year. But then we have Deloitte. They're a global consulting firm.

And their research paints a more cautious picture. They predict that only about 25% of enterprises will be using or planning to use AI agents in 2025, with that number rising to 50% by 2027. That's a pretty big gap between a billion AI agents taking over the world.

and less than half of businesses even dipping their toes in the water. So what's going on here? What's driving this discrepancy? I think this difference in predictions comes down to a few things. On one hand, you have the raw potential of AI agents. They can automate tasks, provide personalized experiences. They can even analyze data and make recommendations.

It's easy to see why someone like Mark Benioff would be so bullish on their future. But on the other hand, there are the realities of business adoption. And that's where those delight numbers start to make sense. AI agents are still a relatively new technology, so companies are going to be cautious. They want to see a clear return on investment before they jump in. And then there's the whole trust factor. Can you really trust an AI agent to make the right call in a complex situation? What if it makes a mistake?

Or even worse, what if it causes some kind of harm? That's where NVIDIA's strategy with those safety microservices comes into play. They're trying to address those concerns head-on.

By building in safeguards, it gets bias, off-topic tangents, even hacking attempts. They're aiming to make AI agents more predictable and reliable for business. Think of it this way. If you're a company and you're thinking about using an AI agent to interact with your customers, wouldn't you feel more comfortable knowing that there are mechanisms in place to prevent things from going sideways? You know that the AI isn't going to suddenly start doing offensive language or making promises it can't keep.

NVIDIA is essentially trying to create a more controlled environment for AI agents to operate in. And that could be a game changer when it comes to convincing businesses to give them a try. Now, even with these safety measures in place, it's important to remember that AI agents are still tools. They're only as good as the data they're trained on, and they still require human oversight and guidance. But what's really fascinating here is that we're seeing a clear shift in focus.

The conversation isn't just about the raw capabilities of AI agents anymore. It's about making them safer, more controllable, and ultimately more trustworthy for businesses. And that brings us back to those Deloitte predictions. Even if they're right and adoption is slower than some of the more optimistic forecasts, it's still significant that they predict half of all businesses will be using or planning to use AI agents within the next few years.

That tells me that this technology is not just a fad, it's here to stay. And as AI agents continue to evolve and become more sophisticated, the questions we're grappling with today about safety, control, and trust

Those are only going to become more important. So what does all of this mean for you? Well, even if you're not a tech CEO or a data scientist, it's worth paying attention to how this AI agent landscape is unfolding. Because if AI agents do become more widely adopted in businesses, it's going to have ripple effects throughout the entire economy. You can change the way we work, the way we interact with companies, even the way we shop and make decisions.

So before we wrap up this deep dive, I want to leave you with something to think about. Imagine a future where AI agents are as common in businesses as like smartphones are today. What are your hopes for their future? What are your concerns? Do you see AI agents as like a force for good, a way to free us from those tedious tasks, at least like new levels of creativity and productivity? Or do you get a sense of unease, like a worry about the potential consequences of handing over so much control to machines?

It's complicated, right? There's no easy answers, but it's definitely a conversation we need to be having, especially as these technologies keep advancing so quickly. I hope this deep dive has given you some new insights into the world of AI agents. And maybe it sparks some questions of your own, because ultimately, the future of AI isn't something that's just going to happen to us. It's something we're all going to shape together. Thanks for joining me for this deep dive into the evolving world of AI agents. Until next time, stay curious and keep exploring.