Eleven years ago, I wrote a blog post with a title that asked the question, “Can you replace a recruiter with an algorithm?”. It provoked an instant reaction and a resounding no from everyone who read it. Fast forward to the present day, and technology has moved on exponentially, but the debate about technology replacing what have always been considered very human aspects of recruiting remains as emotionally charged as it has always been.
The mental models and cultural norms around recruiting run deep, but are we having the right debate? Perhaps the real shift isn’t about AI replacing recruiters. It’s about how quickly recruiters can rethink their role and let go of outdated assumptions. There are several recruiting tasks that AI can already do much better than humans, and the scope is only going to increase.
So what’s left for recruiters? How do we redefine the role of talent acquisition in an AI-driven world? And why is mindset the most critical factor in whether AI becomes a threat or an opportunity?
My guest this week is Nikos Moraitakis, CEO of Workable. In our wide-ranging conversation, we discuss the advancing cognitive ability of AI and how it is changing the way we need to think about what it means to be a recruiter.
In the interview, we discuss:
• The increasing cognitive ability of technology
• Recruiting tasks that AI can already do better than humans
• How AI can do things that are impossible for humans
• Lessons from adjacent areas, such as finance
• The importance of recruiters focusing on high-value tasks
• Agentic AI and Reasoning Models
• What will the ATS of the future be like
Follow this podcast on Apple Podcasts.
Follow this podcast on Spotify.
Matt Alder [00:00:00]:
Will AI replace human recruiters? The debate is fierce and often emotional, but are we actually asking the right question? Keep listening to find out more. Support for this podcast comes from Workable. Workable is the all in one hiring platform that empowers companies of all sizes to attract, evaluate and hire top talent. Effortlessly trusted by over 30,000 companies worldwide, Workable has facilitated over 1.5 million hires, streamlining recruitment with powerful automation and AI driven insights. Hire smarter, faster and with confidence. Learn more at workable.com.
Matt Alder [00:01:02]:
Hi there. Welcome to episode 677 of Recruiting Future with me, Matt Alder. 11 years ago I wrote a blog post with the title that asked the question, can you replace a recruiter with an algorithm? It provoked an instant reaction and a resounding no from everyone who read it. Fast forward to the present day and technology has moved on exponentially. But the debate about technology replacing what’s always been considered very human aspects of recruiting is as emotionally charged as it always has been. The mental models and cultural norms around recruiting run deep. But are we having the right debate? Perhaps the real shift isn’t about AI replacing recruiters. It’s about how quickly recruiters can rethink their role and let go of outdated assumptions. There are a number of recruiting tasks that AI can already do much better than humans and the scope of this is only going to increase. So what’s left for recruiters? How can we redefine the role of talent acquisition in an AI driven world? And why is mindset the most important factor in whether AI becomes a threat or an opportunity? My guest this week is Nikos Moraitakis, CEO of Workable. In our wide ranging conversation, we discuss the advancing cognitive ability of AI and how it’s changing the way we need to think about what it means to be a recruiter. Hi Nikos and welcome to the podcast.
Nikos Moraitakis [00:02:33]:
Hello, nice to meet with you again.
Matt Alder [00:02:35]:
A pleasure to have you back on the show. Please could you introduce yourself and tell everyone what you do well.
Nikos Moraitakis [00:02:42]:
Hello everyone. I’m Nikos Moraitakis. I am the founder and CEO of Workable. Workable is one of the most popular applicant tracking systems and HR systems out there. I’ve been doing this for 13 years now. Excited to talk about recruiting and especially the future of IT with technology.
Matt Alder [00:03:02]:
Fantastic. And let’s start with technology because as you say, you’ve been the founder of a tech business for 13 years. The last sort of few years, we’ve seen this quantum leap forward with generative AI kind of changing everything. What’s that like from your perspective? How is that changing the way that we need to think about technology?
Nikos Moraitakis [00:03:23]:
That’s a good question. Look, the way I think of it at a very high level is that the last decade of software for business was a decade in which we digitized our workflows, basically put everything we do in the real world into a software in a way that the machine can understand what we do and it can visualize it to us in ways that help us as a cognitive tool to understand our work and to move forward with it. But we were very much the pilot. We were executing the process, and we had the system of record that was keeping track of things for us, the next wave, for the next decade with AI, it would be the software actually being able to do some of the things after having observed us for a very long while. How we do things, what kind of outputs do we produce, what decisions do we make? It would be able to, first of all, guide us through the process. Instead of you clicking a button to make something happen, the button will tell you it needs to be clicked. And also, sometimes doing some of the cognitive work we do today, like produce materials, writing, reading things, summarizing it, understanding it.
Matt Alder [00:04:41]:
I guess when we sort of look back at this time, or certainly like the last 10 years or so, we’re going to think that computers used to be kind of pretty stupid, wouldn’t we?
Nikos Moraitakis [00:04:52]:
They were stupid in. In their own very special way. If you wanted to multiply to large numbers, they were a lot smarter than us. But a lot of other things that seem easy to us, they were completely out of their domain. And actually, we all got trained psychologically as a generation that, you know, we have an expectation of what things a computer should be able to do and what it shouldn’t be able to do. What’s it stupid at? The big thing that’s changed that some of the things that previously computers were super stupid at, now they are pretty good at, and we don’t know how to integrate it into our work yet. You see what I’m getting at? I mean, right now, if you have a very large multiplication, you will immediately reach out for the calculator. But there’s many new tasks where we could do that, but we don’t even think of it yet.
Matt Alder [00:05:38]:
Yeah, yeah. So it’s kind of like the, you know, we’re now at a point where the humans are being slow to adapt to what the Computers can do. And the computers are going at a million miles an hour and getting more and more, more and more sophisticated.
Nikos Moraitakis [00:05:50]:
Yeah, I mean, obviously at the end of the day, the computers work for us and we get to define what we want to do and how we want to do it. It’s a new tool that landed, you know, in our lab and now we’re going to have to figure out new ways to use this technology that’s available. And I think that’s very much the job of people like me, the people who, in the software companies in the application layer who are looking at this new computational infrastructure, which is AI, because that’s what it is. Essentially it’s a new computation platform and figuring out what things can we build with that and we can deliver to users so that they can do their job better.
Matt Alder [00:06:34]:
And I think that leads us nicely on to talking about talent acquisition and recruitment and the impact that this is having and this will have. Obviously the big debate, kind of the big picture, picture debate at the moment is whether AI is going to be able to kind of fully replace recruiters or fully replace talent acquisition. It’s kind of much more nuanced than that, isn’t it? What’s your sort of view on that particular discussion?
Nikos Moraitakis [00:06:58]:
Yeah, look, fully replaced is a very strange term. I mean, nothing gets fully replaced. At the end of the day. The reason we do things is because we’re humans and we always want more things to be done and we always want more and we always want other humans to do them for us. To a great extent. However, in the recruiting space, it’s one of the spaces where AI has some very obvious and quick applications. Number one is essentially it’s a job where you get large data sets that describe people in semi structured ways and you need to compare those with a requirement that you have. So the ability to, to march and search through large data sets and find basic level matching of whether somebody has the qualifications for a job is the kind of task that we would look at in the future. In the same way we do large multiplications. I bet you got taught in school how to do it, but you never really do it with a piece of paper because a computer can reliably do it for you. So a lot of the prescreening tasks that we do, a lot of the search tasks that we do on passive candid sets, they lend themselves to AI. We also produce a lot of formulaic material, emails, job descriptions. We, we produce that material in, in a way that, you know, it’s relatively easy for generative AI to be able to do a reasonably good job and save us from a lot of effort. We do a lot of mechanical workflow, applications, scheduling things, reorganizing things, moving things around, reviewing, notifying. That is low skill work that occupies a lot of the time of recruiters and essentially would probably be replaced by agentic or automated AI in the future. That would leave them essentially today. If you ask a recruiter where they spend the time and you add the numbers up, 80% of the things they do are not really hiring and evaluating people and thinking of the strategy of doing that. It’s mostly doing the administrative part of the work. And AI would come through software like workable. So your software will essentially give you these capabilities in much the same way that Excel gave capabilities to accountants that before were unimaginable.
Matt Alder [00:09:31]:
No, actually I think that that’s a really good, really interesting and good analogy with accountants and Excel and Excel kind of really revolutionized what accountants do, didn’t it?
Nikos Moraitakis [00:09:41]:
Yeah, completely. Before accountants were people who could add numbers right in, you know, had good handwriting and put things in columns and add them up correctly. Excel theoretically should have completely obliterated that profession. However, what it did is it led to the financialization of the world because it made it much easier for these accountants to become financial analysts and traders and deal with large data sets and optimize organizations. And now there’s probably more and better paid people working in that field that than before. And I think AI will have a similar effect in most industries. Essentially it’s going to come and do quite a bit of our low skill law firmly like work. It will do some things that are impossible for humans, like comparing data sets of million of things summarizing across large data sets and we will end up doing what we were supposed to do in the first place, which is to meet people, hire people, convince them to work for us, establish a fit, organize our team around the strategy and how to do this.
Matt Alder [00:10:53]:
I completely agree with you and I think that actually it kind of paints a pretty kind of optimistic view of recruitment and talent acquisition in terms of kind of really sort of taking the profession to a higher level and you know, really doing some amazing things. One of the things that strikes me, and it’s kind of come out of a couple of things that you’ve, that you’ve said. We’re not used to computers being this clever, so we’re not quite sure what we hand off to them and what we don’t. And I’ve noticed that there are things that when people Talk about the human aspect of recruiting and what’s really important. They tend to bundle in things like, as you say, matching CVs and those kind of things are something that only humans can do. But as you point out, we’re talking about technology that can compare across millions of data sets in a really short period of time. Do you think that’s kind of holding the thinking back around this, that people are sort of holding on to those ultimately kind of low value tasks that AI could do a lot quicker?
Nikos Moraitakis [00:11:55]:
To a certain extent, it’s natural. And we will all do this during this AI revolution. And there’s a fundamental psychological reason why we would do this is because we ascribe value to ourselves. And except perhaps for the most spiritual of all of us, usually we rationalize this value to ourselves through the mundane things of our everyday life. So we romanticize a little bit what we do if, if these are the five things and do in the office every day and that previously they had, they had to hire someone to do them, you know, it couldn’t be done otherwise. This must be important things that only I can do. Maybe it’s not true for all of them, but you see, you see what I’m getting at?
Matt Alder [00:12:42]:
No, I see exactly what you’re getting.
Nikos Moraitakis [00:12:44]:
We want to believe that there is a special, there’s something special that we add. And the world and the labor market is constructed around this. Our entire life relies on us trading our time for doing things like that to make our living. So we have learned to sometimes, you know, maybe over index on how important they are, or maybe over index on how critical and always right our intervention is. You know the old joke that says that 99% of the drivers say they’re above average drivers. That’s how we do about everything we do at work. And for something, it’s true. And for some things it’s not. When AI comes into some of those tasks, like in the industrial revolution, machines came to replace a lot of things that we used to do. Initially, we would be a bit taken aback. Then it would be obvious to us that that works. You’re not going to dig deeper, faster than a tractor. And at that point we will consider them machine tasks and they would no longer have status and value. And others will. And as humans, we will adapt looking at the things that have status and value, because that’s how we work.
Matt Alder [00:14:03]:
Let’s sort of talk about some of the practicalities of this. So you obviously mentioned writing, job descriptions and matching and the various things that AI can do. Well, what Other things do you think AI can do in recruiting right now? And what might it be able to do in the next sort of, I don’t know, in the next two years or so.
Nikos Moraitakis [00:14:23]:
It doesn’t sound very fancy, but the biggest things it’s going to do over the next two years. We have these reasoning models now that allow you to more easily create reliable, authentic behavior. Which means it’s not just a prompt. You can let it do a piece of work on its own. It has a few steps. The problem with those is that even slight errors in the beginning, 10 steps down could be going crazy. Now with these reasoning models, you can more reliably program some authentic behavior. What this means for us, the user, is that some tasks that we do for the workflow will completely disappear from our UIs. You’re never going to actually engage in the scheduling of something. And you know, right now maybe an AI can do some of those tasks and you click some buttons, but it’s not at the point where you tell it, make sure we see this person like you would do to a personal assistant, so to speak. So we might find ourselves in a situation where your ATS is providing every recruiter with their own PA for every other minor task inside the workflow notification. Chasing up things, cleaning up pipelines, all that good stuff. I think the biggest vector of improvement in the next two years is going to be in that direction because a lot of time get wasted there. Everybody would immediately accept having them done by a computer. The technology is kind of arriving, so I think all vendors will generally go into that direction after the first generative applications.
Matt Alder [00:16:04]:
Yeah, that makes a lot of sense. And this is going to be a kind of this, this is going to feel like a revolution. As you know, computers take over these tasks, do them in a different, more efficient way, changing people’s workflows, opening up the scope to do those higher level tasks. How does that sort of change the skill set that we need in the industry? What do you think the skills of the TA team of the future, or whatever TA is going to be called or wherever it’s going to sit? What skill is going to be needed in recruitment?
Nikos Moraitakis [00:16:38]:
I think for TA teams what’s going to happen is something that already they were trying to do to a great extent to become, to operate a lot more like commercial teams do. I’ll tell you why the bigger impact of this technology in recruiting, if you think about recruiting, it is a problem that is constrained by friction. Friction to apply, friction to schedule, friction to manage big pipelines. You Know, last week I had a user who got 13,000 applicants for a job and they didn’t know what to do. So but theoretically you would want him to get 8 billion applicants for this job. Right. If it wasn’t. If it wasn’t an issue to process that. So the better we become with technology to have wider aids, look at the wider data sets, the employment is going to look a lot more like a real market where everybody can buy everything and everyone knows every price and everybody considers everything. The market by that is not at all symmetrical. It’s very asymmetrical. You know, there’s a lot of things that are unknowns and unknowables. The bigger view on the market that we would be allowed, the more we’re going to have to start operating like salespeople, commercially strategizing, segmenting, going after specific things. Like, it’s going to be a lot more strategic as you work. I think right now we are in a situation where we need to dig the farm and we spend most of our time doing that. But when you have automated machines and you spend a lot of time with what you’re going to plant and what are the seasons and what are the yields and what are the prices, you know, you become a lot more strategic. So I think the moment we get a lot of that menial work, not just not only out, but be able to do it at the scale that was impossible to do before. They would be more like traders and less than diggers.
Matt Alder [00:18:45]:
Yeah. As a final question, and you’ve sort of talked about this all the way through, but perhaps by way of a summary, what does the ATS of the future look like? How are we going to be describing what something like work, workable does in a few years time?
Nikos Moraitakis [00:18:59]:
It’s going to have a lot fewer buttons, I can tell you that. So if you look at them, ats, workable, or any other similar major ats, they all have a lot of screens whose main purpose is to visualize a process for you in a way that you can understand it and operate it. This is what it really is. All the buttons and all the numbers you see over there is a way to visualize the situation for you. And, you know, if it’s a good ats, it will do it in a way that helps you cognitively to actually be more efficient yourself. There will be a lot less to visualize. A lot of the complexity is going to get hidden. A lot of things are going to get automated. So a lot of the parts of the interface will just go away. The other thing I think will happen with a lot of software applications, not just ADSs, is that the UI will be possibly dynamically changing depending on what you’re trying to do. Imagine, you know, like we’ve seen the sci fi movies. They have the spaceship and there’s a holodeck and it shows them the planet or whatever they are about to talk. Whatever they’re talking about, it shows them a visualization specifically made for that. It’s not going to be constrained by one UI that has to have some organization that lets you remember where everything is and understand it. It will make it. It’s going to be like an analyst making you a chart for whatever you need to look at at the moment on the spot. A lot of software is going to become like that. So the software is going to become more like a thing that we go and interact with occasionally to understand what we’re doing, rather than a thing where we spend our whole day inside entering data and updating things.
Matt Alder [00:20:52]:
Nikos, thank you very much for joining me.
Nikos Moraitakis [00:20:55]:
It’s a great pleasure. Always. Thank you very much.
Matt Alder [00:20:58]:
My thanks to Nikos. You can follow this podcast on Apple Podcasts on Spotify, or via your podcasting app of choice. You can search all the past episodes at recruitingfuture.com on that site. You can also subscribe to our weekly newsletter, Recruiting Future Feast. 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.
