The conversation around AI in talent acquisition often focuses on incremental improvements and point solutions – better job descriptions, faster candidate matching, or automated interview scheduling. But this limited view dramatically underestimates the fundamental transformation that’s already underway. While many are still debating whether AI is overhyped, there is growing evidence that it is actually significantly underhyped, with implications far beyond what most TA leaders are preparing for.
So, how can TA move beyond tactical AI implementations to develop the strategic capabilities that will soon be essential?
My guest this week is Jonathan Kestenbaum, Managing Director of Tech Strategy at AMS and a long-time trusted authority in work tech innovation. Jonathan cuts through the noise to explain why current industry discussions around AI are missing the bigger picture – a future where talent acquisition transforms into strategic orchestration and the way organizations attract and deploy talent is fundamentally reshaped.
In the interview, we discuss:
• The unprecedented speed of change compared to previous tech cycles
• The current AI use cases that hint at a much more significant transformation ahead
• Why recruitment processes need a complete redesign, not just automation
• The split between humans and machines and the impact on jobs
• Skills based firing
• Moving from talent acquisition to talent orchestration
• What do TA Leaders need to do right now
• What does the TA Tech stack of the future look like
Follow this podcast in Apple Podcasts.
Follow this podcast on Spotify.
Transcript:
Matt Alder [00:00:00]:
Is the potential of AI in talent acquisition actually being underhyped? Will the real transformation be even more radical than current predictions suggest? Keep listening to find out why. This isn’t just about adding new tools to old processes, it’s about rebuilding the entire concept of talent acquisition from the ground up. Support for this podcast comes from Indeed. They’re a brand that I’m sure you all know as the hiring and matching platform where employers can connect with over 580 million job seeker profiles. But did you also know that their front row seat to the global economy gives them a massive data set which you can access for free? This allows you to see the latest information on job postings, salary trends, and much more. Or did you know that Indeed’s new AI tools make it easier than ever for you to find and connect with active and passive job seekers? There’s much more To Indeed, visit Indeed.com.
Matt Alder [00:01:21]:
Hi there. Welcome to episode 688 of Recruiting Future with me, Matt Alder. The conversation around AI and talent acquisition often focuses on incremental improvements, better job descriptions, faster candidate matching, or automated interview scheduling. But this is a limited view that dramatically underestimates the fundamental transformation that’s already underway. While many are still debating whether AI is overhyped, there is growing evidence that it’s actually significantly underhyped, with implications far beyond what most TA leaders are planning for. So how can TA move beyond tactical AI implementations to develop the strategic capabilities that will soon be essential? My guest this week is Jonathan Kestenbaum, Managing Director of Tech Strategy at AMS and a longtime trusted authority in work tech innovation. In our conversation, Jonathan cuts through the noise to explain why current industry discussions around AI are missing the bigger picture. A future where talent acquisition transforms into strategic orchestration, fundamentally reshaping how organizations attract and deploy talent. Hi Jonathan, and welcome back to the podcast.
Jonathan Kestenbaum [00:02:44]:
Thanks for having me. Excited to be here.
Matt Alder [00:02:47]:
Always a pleasure to talk to you and brilliant to have you on the show again. Just to start off, could you introduce yourself and tell people what you do, what you’re working on? Because it’s kind of particularly important to the conversation we’re about to have.
Jonathan Kestenbaum [00:03:01]:
Absolutely. So I’m on a mission to transform the future of work. I’ve been on that mission for the large majority of my career. I’m a licensed attorney by education, but I’ve been an entrepreneur in the work tech space since I graduated law school I built and sold a work tech company. I then was the co founder of Talent Tech Labs which was recently acquired. And I am in charge of tech strategy at AMS and ultimately spend my days evaluating talent technology solutions and trying to understand how the space is going to evolve and using that information to advise both AMS and our clients on how to build the optimal talent acquisition functions.
Matt Alder [00:03:56]:
And in that role you’re kind of immersed in technology all day, every day, looking at demos, looking at what’s going on with things like AI. What’s your view of the market at the moment? What is going on with AI in TA and HR tech? Sort of where are we?
Jonathan Kestenbaum [00:04:14]:
Yeah, it’s probably the most exciting time that I’ve been in the space. AI has been around for a number of years. It’s not new, but generative AI has brought renewed excitement around AI and it’s a very significant development probably I would say it’s underrated, I think under hyped and gonna make a massive impact on the recruitment space. I always joke that I’ve never really heard a bad sales pitch and they always sound good, they’re always the only, the best, the first, the greatest. And now they all have AI.
Matt Alder [00:04:53]:
Yeah, no, absolutely. It’s kind of interesting and I think to that point I spent a few days last week doing a real kind of deep dive, deep dive into AI. I was kind of at an event with some software engineers and all of this kind of stuff. And I think, you know your comment that it’s, it’s currently underrated, I just couldn’t agree with you. I couldn’t agree with you more. I think that going to change talent acquisition in ways that we can’t even think of at the moment. I mean it’s just, it really is extraordinary. And I suppose to that point, what are the use cases for AI that are really working right now and what do you think? Sort of coming over the hill in the, in the short term.
Jonathan Kestenbaum [00:05:32]:
So there’s a lot to unpack there. But right now we’re exploring things like transcribe and summarize of interviews, job description creation, matching, using image based generation to create job ads. We’re exploring voice based AI agents. I would argue that you’re seeing these microservices at the moment, job description creation, being a microservice and going into 2025 we’re going to see agents and agents that interact with each other. And actually that’s one of my trends that I have highlighted I think will be big in 2025 is the rise of large action models.
Matt Alder [00:06:21]:
Interesting. And tell us a little bit more about that. What are the implications when agents start interacting with each other and things become slightly more autonomous?
Jonathan Kestenbaum [00:06:33]:
I think it’s good to give some framing to that as well. That one of the things that I don’t think people are necessarily thinking about is that AI is slowly eating at tasks, not full jobs. When you think about that microservice, maybe I used to have to create a job description or maybe I would copy it from, indeed from somebody else and make it my own. Now I can use AI to generate that job description. The challenge historically has been when we would bring in, for example, a workflow tool like ATS or CRM. In the past we could do some change management and by the way, companies felt that just doing this much, but we could do some change management and ultimately we’d get our recruiters or our sourcers using these tools and getting the most out of them and we get certain efficiencies, albeit 5, 10, 15%. But now with AI, it’s not just like a tool, it has to be looked at as a strategy because you can’t only do change management. You have to change the actual roles that people do around the technology because it’s eating at tasks. Slowly, agents, to me, it’s starting to do more and more of the work and we have to balance basically the tasks that a human should be doing and the tasks that an AI agent should be doing and a lot of what AMs for example, is working through. And historically, recruitment process outsourcing was a lot about recruitment. People outsourcing. Let us take on the headache of scaling up and scaling down. Let us put those resources in lower cost locations with significant management. And you know, now, now we’re in the orchestration business. It doesn’t matter necessarily how we get that person in that seat, but it matters that we do and that when we do, you know, it’s the right person. And so in that regard, sometimes we’re using recruiters, sometimes we’re using AI agents, sometimes we’re training somebody and placing them.
Matt Alder [00:08:45]:
In a role that makes perfect sense. And where do you think that that kind of split is going to be between what humans do and, and what AI does? Because, you know, I’ve noticed certainly in the last few weeks the debate, for want of a better word, has been hotting up. You know, there are still people claiming that there’ll be no human recruiters by next month. These people have been claiming that for at least a year so far. But at the Same time, there seems to be a lot of people who think that this is just an assistant. It’s got no chance of replacing humans in the recruiting process and all that kind of stuff. And I just think that people aren’t really thinking this through properly in terms of what a machine can do better than. Better than a human. Where do you stand on this particular one?
Jonathan Kestenbaum [00:09:30]:
Yeah, broadly speaking, I think that technology is going to compute and humans are going to engage. That’s a cop out answer. So I’m going to unpack it. I think if we look back at the Industrial revolution, machines were the great equalizer for physical labor and it no longer mattered if you were strong anymore, you know, and was a differentiator for you in a factory. And, you know, you now have the assembly line. And we created basically the assembly line inside of factories and that was meshing humans and machines. Now when that happened, we had at first some job loss, and then long term it created more opportunities. And I would argue the same thing is going to happen with AI, except I don’t know how fast or how, you know, how short or how long that period of job loss is going to happen. I believe that there’s definitely going to be job loss in the short term. I could tell you just examples. I sat on this awesome panel at the LinkedIn Talent Conference where I had Amber Gruel from BCG and an expert from LinkedIn and they were talking about how exactly what I said, recruiters are going to engage and computers are going to compute, and it’s going to be so awesome because we don’t have to do any administrative work. And at the end of the panel, someone in the audience raised his hand and said, hey, that’s all great, but let me tell you how this is playing out in my organization. My CFO called me and he didn’t say, how are you going to retrain all those people that you just replaced with AI and make them salespeople? He called me and said, okay, when am I going to see the EBITDA savings and who you’re going to fire? Right. And so in the short term, there’s going to be change. I think ultimately also certain companies might think that they operate pretty efficiently today using humans, and that could be true. I think that they’ll be able to get higher margins using AI agents in certain parts of the process. And actually there are certain bottlenecks that we have in processes that exist today that maybe AI can help with and streamline processes even further. So I think AI agents, we keep thinking about okay, so it’s obvious to see the tasks of the jobs that they’re eating at and then it’s obvious to see the tasks that humans can now do more of. Maybe we could focus more on building relationships with hiring managers or build a better candidate experience and be more like their agents than just a transactional recruiter that doesn’t get back to them, you know, but it’s not easy yet to see the stuff in between the stuff that wasn’t getting done that if it was getting done, could the process move smoother. The new jobs that are going to be created, some of those are going to be done by humans and some of those are going to be done by agents.
Matt Alder [00:12:32]:
One of the interesting things is a lot of the pushback that I’m seeing about Gen AI in particular is people are using it to write things. All these things sound the same. You can really tell when someone’s using CHAP GPT to write stuff. And I think the, the interesting thing about that and again comes back from a conversation that I had last week when I was doing my AI deep dive. A lot of that is because people don’t know the right tools to use and how to use those tools properly. And actually if people did that, then they, a lot of that, that residence, those objections and those issues wouldn’t, wouldn’t actually exist. So do you, you know, we need a better sort of standard of education and information in this space so people can really appreciate the potential and what’s going on.
Jonathan Kestenbaum [00:13:19]:
Definitely. It also makes me think about what’s happening now with candidates. So, you know, recruiters aren’t the only ones that have access to these tools, so do candidates and they’re leveraging them to apply to jobs. In some cases they’re using them to apply to a thousand jobs at once. In some cases they’re using them to actually take the interview for them. And like, the question is like, is that cheating? And you know, they’re going to have access to these tools on the other side. So, you know, again, AI is like, it’s the great equalizer for knowledge work where essentially what we’re doing now is building the assembly line of knowledge workers. And you know, in that regard, you know, I think it’s interesting to think through, you know, people using these tools. It’s gonna, you know, hurt them in the short term. Like I, I actually don’t have an issue with somebody using an AI to write something for them if, if it’s helping them get their point across more meaningfully. You know, I think if it speeds them up, I mean, it will give them more time to maybe make a phone call and, you know, have a conversation with me. Maybe they use it to strategically think through what they want to say before they call me.
Matt Alder [00:14:33]:
Absolutely, yeah. And I think that’s the thing because I think there’s a bit about people’s concerns that some what have been seen as core skills will disappear because AI can just do all that sort of stuff and is that a bad thing? But actually it’s about developing new skills and being able to do more and better. Better things, isn’t there?
Jonathan Kestenbaum [00:14:50]:
Yeah. And to unpack that, like, I believe that we have to assess people differently. You know, we have to look at agility. You know, I think that’s part of the problem is how we assess people. We need people that know how to leverage AI. I’m actually starting to do some talking about, everyone’s been talking about skills based hiring. What about skills based firing? Because that’s going to be the next five years. We’re going to have skills that are no longer needed in our organization because AI is going to be able to do them and we’re going to have to fire people who only have that skill.
Matt Alder [00:15:26]:
I think that’s a really good point. And I suppose that in some ways technology has been doing this for the last 30, 40 years, that thing in terms of replacing tasks and then eventually replacing, replacing people. But it’s just. This is just a lot quicker, isn’t it?
Jonathan Kestenbaum [00:15:41]:
Definitely. The speed is insane. I mean, it’s wild to think about how much things have changed from a technology perspective with AI over the last two years Now, I would argue adoption is still behind the curve. I mean, even at AMF, if we look at our client base, only 50% of our clients have policies that allow them to use AI. And so AI adoption has been slower. And it’s also like there’s so many layers of complexity that go into it. I mean, think about it. The average person in a talent acquisition environment uses eight to 12 systems. You’re going to have now eight different siloed data sets. Each of those technologies is going to bring in their own AI agents, like getting the process to move smoothly across that data. And data super important to AI is going to be super challenging. So these are some of the things that at ams we’re shooting to solve is how do you aggregate and centralize that data so that you can apply AI agents across the stack?
Matt Alder [00:16:44]:
Yeah. And I suppose that from a kind of a future skill perspective within ta, if you’re a, you know, if you’re a TA leader or an rpo, having that, the knowledge and skill to have that sophisticated approach is going to be, is going to be critical. And I don’t think that there are many people out there who are currently able to, you know, to do that, who are in that, that kind of role.
Jonathan Kestenbaum [00:17:06]:
Yeah, we are moving from talent acquisition to talent orchestration. And the talent acquisition leaders of the past are going to become talent architects.
Matt Alder [00:17:17]:
No. Yeah, 100%. And what would your advice be to, I suppose both to kind of, you know, employers sort of in a collective, but also, you know, the individuals working in talent acquisition, what should they be doing right now?
Jonathan Kestenbaum [00:17:30]:
I think they need to stay on top of, of the technology and the tools that are out there that can make them more efficient and become proficient in them. I think that there’s, especially in the us, the uk, the eu, there’s compliance issues around leveraging these technologies that have to be dealt with and getting training in how to leverage these technologies and use them compliantly. For example, everyone always thinks about compliance issues with AI being around bias. But you know what, bias is really a pretty binary problem to solve. It either is biased or isn’t. And there’s tests to test that. That’s only 10% of the challenge. 90% of the challenge is how you use these technologies in your process. And for example, if I was to use a social sourcing tool, create a list of diverse candidates, and then rely upon that list to a hire, I violated US Federal civil rights laws. And that’s the thing about training people how to use AI. And there’s not a lot of training yet. And it’s a new skill. You know, it’s a, it’s a new muscle that organizations have to build from.
Matt Alder [00:18:42]:
The, all of the, the tools and the use cases and the technologies that you’ve seen over the last sort of two years, what’s the one that surprised you the most? What’s the one that really you did not expect AI to be able to do?
Jonathan Kestenbaum [00:18:56]:
Yeah, I mean, I am fascinated by voice agents. I mean, I think the quality has gotten so fantastically good that I enjoy having a conversation with them. I’ve made a deep fake of myself. It’s, it’s scary and it just makes me think that identity is, is gonna be an issue moving forward. You know, I think blockchain makes a comeback in solving the identity problem because it’s gonna be really hard to differentiate what’s real and what’s not. I, I, I think, and by the way it’s, it’s, it’s not just voice generation that’s interesting. It’s video based generation. It’s, you know, image based generation. And these things are just insane.
Matt Alder [00:19:39]:
Yeah, they are. I, I cloned my voice using AI a few weeks ago and yeah, I kind of didn’t even, I didn’t even do it properly because you’re supposed to. The tool I was using is supposed to upload five hours worth of you speaking, which is a podcast that I have. But I, I just uploaded half of it was terrifying. It sounded just like me.
Jonathan Kestenbaum [00:20:00]:
I just read an article about some, somebody who had, you know, got extorted. Their parents got a call that they were in trouble and they got extorted by some criminals overseas. And the way they got that person’s voice was through their Instagram reels.
Matt Alder [00:20:18]:
Wow. Yeah, I already had to, I’ve already had to tell quite a few people that I’ve sent it to people and said, this is a clone of me. So this is, this is what’s. So, it’s. Yeah, it’s fantastic and terrifying all at the same time. As a final question, what do you think the tech stack of the future looks like from a TA perspective?
Jonathan Kestenbaum [00:20:41]:
It’s a really good question. And I think there’s going to be three big pieces of technology that don’t exist today that will exist in the future. I broadly think that the UI layer is going to be obfuscated into a chat based interface that exists within the flow of work and many of these workflow engines will kind of exist within that, which I think is highly disruptive to some of the existing players with significant market share. In the short term though, I believe you’re going to need to have an integration layer that integrates into all these point solutions to centralized data. I believe that you’re going to need an orchestration layer to orchestrate tasks between humans and AI agents. And I believe that you’re going to need an enterprise agent platform so you can deploy agents across your stack and across your process. And as a result, I think that there’s going to be two kinds of companies. There’s going to be companies that either use a core system like a workday, a, you know, a servicenow, one of these big core platforms. And I think you’re gonna see more people moving in that direction, even if their features in specific niche areas aren’t as good because data centralized and it’ll be easier to apply AI. And then I think you’ll see organizations that still want best in class tools. Get to work with an organization like AMS who can help them centralize that data and apply AI to across the data set.
Matt Alder [00:22:17]:
Jonathan, thank you so much for talking to me.
Jonathan Kestenbaum [00:22:21]:
Thanks for having me.
Matt Alder [00:22:23]:
My thanks to Jonathan. You can follow this podcast on Apple Podcasts, on Spotify or wherever you get your podcasts. You can search all the past episodes@recruiting future.com on that site. You can also subscribe to our weekly newsletter, Recruiting Future Feast and get the inside track on everything that’s coming up on the show. Thanks very much for listening. I’ll be back next time and I hope you’ll join me.
