Josh Bersin: Why we’ll learn to love AI agents

Few analysts, executives, or academics have studied the intersection of technology and the workplace more deeply than Josh Bersin. The founder of Bersin & Associates, a top research and advisory firm focused on HR technology and talent management, Bersin has spent over two decades tracking how digital tools are reshaping the workplace—and now work itself.

One new area of focus for Bersin is AI agents—next-generation AI tools that can multitask and automate in concert with other agents and bots to get things done. In a recent interview with The Works, Bersin described how he sees AI agents evolving into a roster of invaluable digital coworkers. 

How quickly do you see AI agents emerging in different sectors and business functions?

I think there's going to be a very large market for expert agents that emerges in different domains. It’s just starting to take off now. 

A lot of AI agents out there now are what I would call information or knowledge agents—they can answer questions, find things, and generate content. But what's coming is a much larger market for expert agents in specific domains. You’ll have agents for customer service, lawyers, doctors, HR roles, and so on. The agentic part of it—where the agent pulls off many tasks—is what's next.

I think it’s going to feel like science fiction at first, but it’ll quickly become the norm. AI agents will be embedded across all business functions—customer service, HR, legal, finance, you name it. They’ll streamline processes and allow companies to operate with fewer resources while improving speed and efficiency.

We’re just at the start of this revolution, and as these systems get smarter, they’ll fundamentally change how we work.

What are some enterprise use cases for AI agents that have big potential?  

In corporate training, there's a $300 billion or so market that includes everything from education to onboarding, upskilling, certification, and leadership development. Traditionally, instructional designers and HR professionals spend months developing courses. You interview experts, create content, and load it into learning platforms.

Now, there are agents that automate much of this. They can gather information from experts, generate the course, send it out for review, and launch it to the right employees. This kind of automation cuts development time from months to days, and the agent even follows up with learners to make sure they complete the training.

There are hundreds of steps like this that HR people are doing that can be completely automated. The same thing is true with recruiting tasks. Every company's process is different, but if you had a tool, you could map your processes into your roles. Setting up AI agents may end up being a little bit like writing an Excel macro.

Where else will AI agents make an impact?

I think we're going to see them everywhere. Customer service is a big one. Right now, agents are mostly automating tasks like sending emails or scheduling meetings, but they're evolving into what I’d call digital employees.

Take recruiting, for example. We’ve got agents now that can create job postings, review applications, rank candidates, and even send customized emails to applicants. These systems are becoming more intelligent and taking on more complex tasks, like assessing cultural fit or asking nuanced questions during the hiring process.

So, I actually think AI agents will be a little more like a digital employee than you might imagine. Not everybody likes thinking about it that way, but I don't think it's going to take very long before these agents are really like hard-working employees that just don't sleep.

What does that mean for job replacement, especially in administrative roles?

I don’t think it’s going to lead to mass layoffs, but roles will definitely change. For example, in recruiting, there’s a role called a recruitment coordinator. These people handle tasks like scheduling interviews, following up with candidates, and sending emails. A lot of that is going to be automated, but it doesn’t mean those people will lose their jobs. They could move into more strategic roles or help train the AI systems.

Companies I’ve talked to don’t want to fire people—they want to redeploy them into other jobs. But I do think, over time, companies will get smaller as administrative tasks become more automated.

It’s easy to see the upside potential, but what challenges and problems do AI agents present?

Data privacy is a huge concern. AI agents are pulling data from all sorts of sources, and the risk is that they access information they shouldn’t. An agent trained on salary data could give an applicant insight into pay ranges that the company didn’t intend to disclose.

Once an agent accesses certain data, it can retain that information and use it in ways you didn’t expect. So there’s a lot of work to do on data security and setting clear boundaries on what agents can and cannot access. Vendors are building tools to address this, but it’s definitely an area that needs close attention.

What other workplace applications for AI are you excited about?

AI coaches are an emerging area that I think will really take off. We’re seeing companies develop AI-based systems that can provide coaching, feedback, and support to managers and employees.

These systems can help with things like leadership development, giving tips on how to handle certain situations, or even just offering encouragement. There’s a lot of research behind how AI can engage with people on a personal level. In the business world, this could significantly reduce the cost of coaching, which can run anywhere from $500 to $1,000 an hour.