Most of the people talking about where AI is taking hiring come from one of two camps. Some started software companies and aren't of this world. The others are staffing people trying to talk AI.

I live in both worlds. I've spent most of my career in recruiting and I built Aeon Hire, and my company has done 15,000 corporate hires across 43 countries.

So here's my view on agentic AI in recruitment and hiring.

There are many things that should be automated, and agentic AI is going to get rid of a lot of the work recruiters hate. But some people get confused and want to automate things that shouldn't be, and what's funny is those people are never from our industry.

Hiring is a human decision, and it shouldn't be automated away.

The Short Version

• Agentic AI in recruitment is specialist AI, with agents built for specific areas that work together to get outcomes.
• When we did research back in 2011, the time spent looking for candidates versus talking to them was a 60/40 split, with the 60 going to looking.
• Agents should take that data janitorial work so recruiters can build rapport and hiring managers can make the decision.
• AI has broken trust between candidates and hiring managers, and I think validated information on both sides can rebuild it.

What Agentic AI in Recruitment Means

Agentic AI is specialist AI, where each agent has specific areas of specialization, and you can deploy those agents in multiple different ways so they work together to get outcomes. People get confused and think it's just setting up a rule.

They have reason to be confused. In 2025, Gartner estimated that only about 130 of the thousands of companies selling agentic AI offer the real thing, and it calls the habit of relabeling chatbots and old automation tools agent washing.

So what does specialist look like in hiring? What if you had AI agents sourcing on your healthcare roles, your technology roles and your finance roles, at big enterprise companies and at small and medium companies, in South Florida or New York City or Seattle? Those are the kinds of areas where an agent can have expertise of its own and work alongside other agents.

Agentic AI Should Take the Data Janitorial Work

Look back to the days of Mad Men in the 60s. The executive assistants were in the room with the CEO and the CFO typing up the notes, and they sent those notes to the team, and that's how everything got recorded and documented.

What happened next is the computer, the voice recorder and eventually Zoom came into play, and nobody needed to sit there with a typewriter anymore. Going back to that today would seem pretty rudimentary.

Did that change the fact that there were executive assistants? No. Did that make the old technology obsolete, with new technology doing things better and faster? Absolutely.

Did anybody get upset that they couldn't do their favorite thing, typing? No, because so many parts of everybody's job are administrative. It's data janitorial work, the type of work you don't want to do, and everybody's got it in every job.

I really believe the things you hate to do are mundane or formulaic, and they're not the things that make you use your brain, challenge you or make you be strategic. Now imagine getting rid of all of that, because that's what agentic AI is going to be able to do.

Generative AI is already saving recruiting teams time. LinkedIn's Future of Recruiting 2025 report found talent acquisition pros who are experimenting with or integrating generative AI save about 20 percent of their work week on average, which works out to a full workday.

Looking for Candidates Is the Part Nobody Enjoys

Looking for candidates is one of the most arduous parts of any recruiter's job. When we did research back in 2011 on how much time was spent looking for candidates versus talking to candidates, it was a 60/40 split, and the 60 went to looking.

It's not something anybody really enjoys. Some people are really good at it, but it's not the thing you want to be doing if you're in recruiting.

That's the example I use for agents, and they can be built to do more than identify candidates.

1. Identify the candidates for each role and market.
2. Enter them into your system.
3. Maybe initially pre-screen and engage them.
4. Set you up as the human to do the work you love.

That last step is the reason for the other three.

The Work Agentic AI Should Hand Back to People

The work you love is building a rapport, learning about people and their motivations and helping them understand if you have a role that's a fit for them, a company that's a fit for them and a manager that's a fit for them.

That's the thing that excites you. It's so gratifying when you help that person get a job, and it's so gratifying for the candidate when they find a job that gives them purpose instead of something they don't enjoy doing day-to-day.

And when you're the hiring manager and you hire the person who's going to take your team from a B to an A, it feels so fulfilling.

Why the Wrong People Keep Trying to Automate Hiring

Everybody struggles with hiring, and a lot of people don't enjoy it. It's a necessary evil that takes them away from their core day-to-day job, which is why there's little prep, there's little post-work and most people aren't that great at it, at least when you look at the data from a hiring perspective.

So whenever you have something people don't love to do, what do software people like to do? Let's automate it, let's use AI and let's not waste our time on it.

That's wrong, and it's wrong because you have the wrong people trying to solve the problem, people who think hiring sucks and want to automate it. I know better, and I know that recruiting and hiring in particular are human-centric.

Hiring Is a Human Decision and It Belongs in the Room

As long as people are working together for any type of common good or cause, it's always going to be better to meet that person and understand them face to face. Just because COVID happened and Zoom happened, and we can do everything remote nowadays, doesn't mean that we should.

I don't believe hiring should be done 100 percent remote or 100 percent automated. It's a uniquely human experience where trust is built, chemistry is built and understanding is built, and then you make a really big decision.

Think about it from the candidate's side. Where am I going to spend my 40-plus hours a week? Who am I going to be working with? What am I going to be working on, and what is it going to be about? That's a really important decision for anybody to make.

Google, Cisco and McKinsey brought back face-to-face interviews for at least part of their hiring to counter AI cheating, Axios reported in August 2025.

The human part also means picking up the phone. When I give an internal talent acquisition team the perfect resume and the cell phone number of the exact person they want, I still hear "we can't get a hold of them" or "they didn't reply to my email." What are you talking about? You've got to get them on the phone. One of the reasons MSH's talent acquisition work is successful is that we train our people to do the warming up and the calling.

How AI Broke Trust in Hiring

The problem with AI right now, in the hiring industry in particular, is that it's only made things worse. Everybody is more distrustful than ever.

Half of people don't even know if the role they're applying for is a real one. That's what a 2025 Gartner survey of job candidates found, and Gartner also predicts one in four candidate profiles worldwide will be fake by 2028.

There's also a ton of distrust on the manager and interviewer side. They're all looking at things and saying, "Is this the real person? Are they using AI to cheat? Is their resume even real?"

So now it's AI fighting AI. Resumes get put together by AI and shotgunned out to a thousand jobs, then AI reads them and shortlists or down-lists people, and some companies have people doing AI avatar screening on top of that.

Something that was already kind of an art and a dance, where you weren't really sure what you were walking into, has become even more distrusted and more disingenuous because of AI.

Greenhouse heard it from job seekers too. In its 2025 AI in Hiring Report, 46 percent of US job seekers said their trust in hiring had dropped over the past year, with 42 percent blaming AI directly.

How AI Can Help Rebuild Trust in Hiring

To me, this is just a temporary thing. The industry is still figuring out how to best use AI in this space, and while AI has broken trust down now, I think it can rebuild it in the future.

That's where we're looking at Aeon Hire, the screening and evaluation platform I built, to take things to another level. Right now interviewing is mostly gut instinct.

The idea is validated information and an understanding of what matters, with candidates putting in their own behavioral profile and the platform learning about the managers and the teams too, so both sides can make decisions with validated information.

Think about the credit score in lending. The FICO score didn't take the place of all lending, and it didn't make all lending perfect. It made lending more predictable, and that's what validated information can do for hiring.

The goal is to get away from this lack of trust and build a trust network. And it still comes back to the root of the problem, because interviewing for a job and making a hire are really, really important to do in person.

How I'm Rolling Out Agentic AI Inside MSH

Inside MSH, we're training the whole team on AI and rebuilding the admin work we used to do by hand into agentic workflows. And I'm not going to fire anybody to do it.

It's the same playbook in every function.

1. Hire people who are innovative and curious. Most of my team already knew our industry was changing and wanted to be at the front of it, and that's a big part of why they work here.
2. Train in groups. If you're in recruiting, everybody in recruiting should be building the same skills, and it shouldn't be completely bottoms up.
3. Invest in your people. You can't play with scared money, and learning and development is one of the key aspects of retention.
4. Hold them accountable. What skill did you build? What tokens are you using? How are you using this to make things better?

This is part of the job now, and it's a KPI. If somebody isn't building, that's a red flag you work through with them.

Questions I Get About Agentic AI in Recruitment and Hiring

How should you use agentic AI in recruitment?

The example I use is looking for candidates, because it's the grunt work nobody enjoys. Picture one agent working your healthcare roles in South Florida and another working your technology roles in Seattle, so your recruiters spend their days talking to people instead of hunting for them.

Will agentic AI replace recruiters?

I'm not going to pretend there won't be disruption. My belief is that over the next two to five years, we're going to walk into the biggest dislocation of human labor, because a lot of people won't have the skills the jobs of tomorrow need. But look at travel agents, who you would have thought were done once the internet took off in the mid 90s to early 2000s. They're still around, focused on luxury white glove service, because the job just shifted what became the priority. When I look at recruiting, I still think there's a very, very powerful aspect of it that's human to human.

Can candidates use AI to apply for jobs?

Plenty already use AI to apply. In the same 2025 Gartner research, 39 percent of candidates said they used AI during the application process. That's the candidate half of AI fighting AI, and it's why validated information has to work for both sides.

Where I Land on Agentic AI in Recruitment

There are absolutely things agents can do to make sure you're spending your time on the hiring processes that are most important and that lead to great decisions. But it shouldn't be fully automated away.

You can use AI agents, but this industry needs to rebuild the trust. It has to do right by the candidates and the managers who are making these very, very important decisions.