How Companies Are Already Using AI Agents
12:40 Share

ADP knows any big thing, any small thing, any trendy thing, even a trendy thing that everyone knows isn't a great idea, but management just wants us to give it a try for a bit can change the world of work. From HR to payroll, ADP designs forward-thinking solutions to take on the next anything. Welcome to Tech News Briefing. It's Monday, January the 13th. I'm James Rundle for The Wall Street Journal. Self-driving cars can perform well in ideal conditions, such as the sunny streets of Phoenix. But when it comes to ice and snow, they can struggle.

We'll hear why one company thinks smarter tires may be the solution. And then, generative artificial intelligence is being rolled out in earnest by major companies, covering everything from financial analysis to marketing and customer service. WSJ reporter Bell Lin joins us to talk about how some companies are using the tech right now to save money and in some cases replace humans.

But first, while autonomous vehicles, or AVs, are now common sights in the Bay Area, they haven't seen the same success in colder parts of the US. That's because, among other reasons, AVs often struggle with understanding the particulars of driving in snowy or icy conditions. WSJ reporter Isabel Busquets joins us to talk about how tiremaker Goodyear is developing new tech aimed at solving these problems.

So Isabel, what has been the issue with AVs and how they handle conditions like cold and ice? One of the biggest problems is actually the braking system. So cars take longer to brake when there's inclement weather, when there's snow or ice or wet streets. And that's one of the things that self-driving cars have struggled to account for so far. And they don't always know the conditions of the tires or what type of tires they have. And that also affects

how long it will take to break in various conditions. So those braking systems have sort of been a sticking point. So why is this becoming an important issue for AVs now? It's becoming an important issue because we're sort of at the point where they're scaling up in terms of the number of passengers they're taking. Waymo is becoming pretty popular out on the West Coast in cities like San Francisco. As of October, it was doing 100,000 paid rides a week.

We're seeing other companies like Tesla and Amazon trying to move into this space as well. But we're seeing it a lot in these sort of drier West Coast climates. There's the expectation that these self-driving fleets are going to want to scale into more inclement weather, tougher climates at some point. And so this is definitely an issue that they'll have to solve.

What has Goodyear done to try to break through this challenge? Yeah, essentially what they've done is they've created this technology that uses data about what the weather is and also data about the type and condition of the tire to create a more nuanced and accurate prediction of how long it will take a given car to break in

in given conditions. So it can grab data about the weather either off the front and rear cameras on the vehicle. It can also tap into third party weather data sources. And then, of course, it knows what kind of tire it is and how old that tire is.

And then based on that, it can generate a more accurate prediction. They produce these sensors that they can put in tires and that would probably give them the most accurate prediction of exactly what the friction level is on the road surface at that given moment. Because that's really like what they're trying to get out with all that data about the weather is like really what is the friction on the road.

but that's something that's still emerging. They haven't really found a great way to power those sensors yet. They're looking at batteries and other kinds of things.

It's not just AVs that have issues with braking distances. Is this just applicable to AVs or can human operated cars benefit as well? That's a really good point. Actually, there's a mandate from the National Highway Traffic Safety Administration that by 2029, all vehicles will have to have these automated emergency braking systems. So that means that any time any vehicle, whether it's human operated or

or self-driving headed for a pedestrian, an obstacle, another car, it has to have an automated braking system. So this is something that could really be applicable to those types of cars as well as AVs. And Goodyear is hoping that that 2029 mandate will spur adoption of this kind of technology.

That was our reporter, Isabel Busquets. Coming up, major companies are now deploying AI agents in a number of roles. Could they eventually displace people? And has it happened already? That's after the break. ADP imagines a world of work where smart machines become too smart. Copier, I need 15 copies of this. Printing. By the way, irregardless, not a word, Janet. Yeah, I know. Page six should be regardless of or irrespective of. Just print them, please.

If it were a word, Janet, it would mean without irregard, which is... Copier! Switch to silent mode. Let's put a pin in it. Anything can change the world of work. From HR to payroll, ADP helps businesses take on the next anything. Generative AI's ability to process vast amounts of information and talk conversationally has prompted some companies to deploy it in specialized roles.

At ratings agency Moody's, for instance, AI agents take on some of the work traditionally done by analysts, while eBay uses AI to write compelling listings copy. Some companies are even starting to refer to these bots as employees. Despite this, the technology still requires a lot of oversight. WSJ reporter Bell Lin joins us now with more on how companies are embracing AI.

Bell, you looked at how several companies are deploying AI. What was your overall feel for how the technology is being used? AI agents are relatively new. They're certainly pretty immature in terms of overall company adoption, but they are promised to really deliver this next generation.

evolution of productivity. Because right now, when it comes to AI's use in companies, really across the whole range of sectors, there's not a lot of productivity boost or top line or bottom line impact that they're having. But AI agents are different because they actually deliver autonomous work, supposedly. So if they work, and some of these companies that I talked to say that they do work as promised,

They can actually do work on their own. They can even interact with other agents to ultimately make decisions and make things a lot easier for their human counterparts. You talked to several companies, including Johnson & Johnson, Moody's and eBay.

Were there any examples that stood out to you in particular? Yeah, I'm really excited about this idea of multi-agent systems. And that's where you have multiple agents working together. And Moody's has 35 agents that they've built. And these agents are all designed with different personalities, with different job tasks. And so for what Moody's does, which is conduct financial research on behalf of their clients,

What they told me is that it's really important for each of these agents to have distinct roles because you want them to disagree on, say, should I invest in this company? Because one agent might be looking at something specific like geopolitical risk and other agents just looking at the P&L statement and saying things look really good.

And so you want them to have these specific avenues that they're looking down and saying, how do I weigh these different factors and tell the client the best answer on whether or not to invest in a company? The Moody's example is really interesting. We've heard companies say that AI is used to augment humans a lot of the time. But looking at some of the use cases in your story, it does seem that humans are being replaced in some instances.

Did you get that sense? Oh, absolutely. None of what AI does and the sort of next wave of automation is possible without some human replacement and some, you could call it, augmentation. The idea is that a lot of these agents or chatbots can do this work for humans a lot better, but that means that there's less work to go around. The best case scenario is that

the agents help us do our jobs better and faster. And then the worst case scenario is that we don't need as many of us because there's more of our digital colleagues. What workers should be concerned then?

Workers should certainly be concerned when they feel like they're not able to learn or use AI to the best of their ability to stay up to date on how the technology is evolving. What I often hear from whether it's economists or people who followed these job trends is people should be keeping up with what AI can do and harnessing it so that it's not really the people who will be replaced by AI, but it's the people who don't use AI who might be replaced by it.

From your reporting, how are companies controlling the deployment of AI in terms of oversight, in terms of monitoring? Some of these AI agents, and really a lot of them, are not quite ready to run amok. With more autonomy comes more challenges and more cybersecurity risk.

So the great thing about an agent that can tell you whether or not to invest in a company or submit a holiday request for you is that you don't have to do the work. But at the same time, that opens up a whole potential world of pain because you don't want that agent to access data that it shouldn't have about your colleagues or about pay information from the rest of the company that could cause it to hallucinate and give you the wrong investment. And so companies are really approaching agents

agent rollout with caution and trying to even programatize that so it doesn't have to be so manual. But right now it's important for the manual processes to be in place.

And did any of the companies talk about how these AI agents are maybe saving them money or are they costing them? Yeah, absolutely. The question around AI's return on investment is a really big one right now. Among the companies that we talked to, there's been an increasing sort of scrutiny on whether or not these investments can pay off. And that's why AI agents are sort of looked to as investors.

maybe a promised answer that they can lead companies into the promised land of investment returns. Right now, that's not quite proven out, but we are hearing more that companies can do things like reduce the number of workers that they might plan to hire. So that's in the instance of Cosentino, a company that specializes in creating these surface materials for interiors and exteriors. And they actually automated away a lot of their customer service processes.

And so that's a true cost savings. And so for them, AI is a really good return on investment. That was our reporter, Bell Lin. And that's it for Tech News Briefing. Today's show was produced by Julie Chiang with supervising producer Catherine Millsop. I'm James Rundle for The Wall Street Journal. We'll be back this afternoon with TMB Tech Minutes. Thanks for listening.

ADP knows any big thing, any small thing, any trendy thing, even a trendy thing that everyone knows isn't a great idea, but management just wants us to give it a try for a bit can change the world of work. From HR to payroll, ADP designs forward-thinking solutions to take on the next anything.