The Real Value Of AI Agents

If you believe the hype, AI agents are everywhere at the moment in teleacquisition. However, when you dig a bit deeper, it becomes clear that many vendors are simply rebadging existing chatports and assistants. Real agents handle entire workflows autonomously. How are they being used in TA and what value are they creating? Keep listening, to find out. Support for this podcast comes from Talent Pilot, the first end-to-end AI-native platform for recruiting. Let's be honest, no one becomes a recruiter to copy and paste CVs, chase feedback, or write interview notes. Talent Pilot is helping the role of the recruiter to quickly evolve from a junior admin-heavy one to one of the most strategic roles in the entire organisation. With Talent Pilot, recruiters can build their own hiring workflows by deploying AI agents at different touchpoints of the recruiting process, to source, screen, and select the best talent. They're offering a free demo for everyone who listens to recruiting future, so head over to talentpilot.com/mat to experience the new era of recruiting. That's talentpilot.com/mat and it's Matt, M-A-T-T. There's been more of scientific discovery, more of technical advancement and material progress in your lifetime of a month, than all the ages of history. Hi there, welcome to episode 748, a recruiting feature with me, Matt Alder. The AI agent marketplace has become a confusing landscape that's full of chat bots and co-pilots that aren't agents, claiming revolutionary capabilities. But genuine AI agents represent something fundamentally different. They're digital workers that can handle complex multi-step processes independently, making decisions and adjustments along the way. The technology is already here and working, and the employers who are succeeding with agents are focusing on change management, not just technology deployment. So, what are the early results looking like? And how will a Genetic AI change recruiting in the months and years to come? My guess this week is Tom Zabecki, founder and CEO of talentpilot. In our conversation, Tom shares case studies demonstrating how AI agents are reshaping recruiting workflows, and we discuss what autonomous hiring is going to look like. Hi Tom, and welcome to the podcast. Hi Matt, pleasure to be here. Thanks for having me. It's absolutely pleasure to have you on the show. Please could you introduce yourself and tell us what you do? Sure, my name is Tom Zabecki. I'm the founder and CEO of talent pilot, where we are building an agentic talent acquisition system that basically helps you source, screen, and select the best talent. And what actually means is that you just tell the system we were looking for, let's say a marketing expert with at least five years of experience in Google AdWords, Living in London, and our AI agents go through the internet, you know, find the right people for you, reach out to them, interview them, and then schedule the interview with the candidate direct material calendar. So this is how it works. Fantastic, and let's just start with almost with some definitions around a Genetic because every company seems to have an agent, or claims to have an agent these days. And to me, some of these agents just look like rebadged search engines or glorified chatbots. What's the difference between, you know, an agent and a chatbot? What does an agent do that's different from some of the other things I mentioned there? Yes. So from the broader perspective, both AI agents and AI chatbots fall under the umbrella of artificial intelligence systems, but they are different in the level of autonomy they can work, which means that for example, AI chatbots can work autonomously for a couple of minutes. Example is, you know, JGPT, and it's more to deep research, where you ask the JGPT to do a very, very deep analysis on the topic, it does it, and then provides you with the outcomes, usually within a couple of minutes. And AI agents, they are having much longer autonomic windows, which means that the best agents on the market currently can work autonomously for up to 90 minutes, which is, you know, way more than the traditional tentbots. That has really interesting. I've not really heard it described like that before, so that kind of makes a lot of sense. And what does that mean in terms of their capability, you know, what are they able to do? I suppose, particularly in the context of italic position, but also more broadly. You know, if you ask 10 different people what an AI agent is, you probably get 10 different opinions. But in my perspective, an AI agent is actually a digital worker that can do some work on its own and to end and deliver the outcomes for you. And in the sense of recruitment or talent pilot, for example, we have this AI agent that can do interviews with candidates. And this agent can conduct the AI interview for up to 60 minutes. So it's very, very autonomous agent that can talk to the candidate, ask him or her questions or answer any question the candidate might have for really 60 minutes. But the real beauty with AI agents is if you put more agents, after each other. And you will build something that's called agentic workflow to cover the whole process. So in our specific case, and this is what I mentioned at the beginning, what we do is that we have several agents that can cover the process and to end. We have an agent that can go to the internet, to LinkedIn, to find the relevant people based on your request. Then if it finds them, it can hand them over to our outreach agent that based on the data about the people, about the position the company can really create personalized outreach message to those candidates. And for those candidates that reply and are interested in the opportunity, there is the next agent that is the interview agent that can call the candidate, talk to them, explain them the opportunity. And again, if the candidate is interested to kind of go more in the process, go further in the process, then there's the last agent, which is the scheduling agent that tags these kind of interested candidates and schedule the meeting directly to the reporter's calendar. So it's not just like one agent, you know, that the beauty is that you put different agents at work. So when you do that, you start the conversation, you still describe that some of what we're put in. I'm looking for for this type of person with these type of skills. And then they delivered it basically a pre-vetted short list of people that they can take to a human interview. What kind of decisions are the agents making along that workflow to be able to get to that point? As you can imagine, there is a lot of regulation popping up recently. There are many new legislations written in the US specifically. There is this new Colorado AI Ag. There is the Illinois AI Ag. We all know the New York, local law 144, or the infamous EU AI Act. And we are trying to build or we are not trying. We are actually building what is called responsible AI. So we know that we are the language most sensitive data that there is, meaning data about people, employees, companies, and so on. And we need to make sure that the AI works properly with this type of data. The AI does minor decisions in the whole process, but never the big ones, like whom to hire or anything like that. This is one thing. But second thing, what we also need to embed into our system is full traceability of the AI reasoning. Therefore, the user, if it's a recruiter or someone else from the company, can always ask the AI how the AI reached the outcomes it did to better understand the flow of the AI. So this is obviously a very critical principle and the developed plan of AI for recruitment and employment is this human in the loop, full traceability of data, you know, documentation and measuring or auditing the biases and fairness of the system itself. So with the clients that you're working with, what have they learned, what have you learned so far using this kind of approach? So I think the biggest lesson learned I have from the last year or so is that it's not just about AI transformation itself in companies. It's more about changing management transformation because you know, we have the technology. The technology is here. The tools already, some companies are using it, some don't. And the real difference is how you use a company approach change management people because people don't like change. Let's be honest and AI from my own perspective will have very, very big change in our day-to-day workflows at work and it's scary. It's honestly very scary because it can really put our world upside down and I can tell it from our own experience because what I described basically does the whole hiring process on its own and when are the recruiters and was the role of a recruiter. So it's scary but it's unavoidable. We are going there no matter what, we all know that. So this was kind of the key lesson learned is that you need to focus on the change management part. And if we are in our case, if we are starting new cooperation with customers and we are working with the biggest banks, biggest technology companies in the world and so on. So we have this real information and this real data. When we are starting with those types of companies or any type of company, we usually start with scanning the organization to identify the open-minded for the thinking, innovation-driven people that might become or they will become ambassadors and they will bring the change to the organization. So this is the most crucial part is to have someone who really wants to do it, who is open to it, who is open to new experiences, who is open to change their workflows because they know the technology shift is coming. And once we have identified these people, we are usually starting smaller. We are starting with actually this group of people. We are not implementing it across the whole organization because it might die out. So once we have those three, four, five, six, seven, ten kind of innovators within the company, we are defining the process itself and how we want to tackle the process because every organization has slightly different recruitment approach, slightly different hiring process and Talent Pilot as a product is very, I would say, customizable. It's very flexible. We have many agents for different tasks in the hiring process and it's up to the organization if they want to use all the agents and cover the whole process or they just want to use one specific agent to cover one part of the process. For example, you know that scheduling is very demanding tasks in the in the recruitment process. So the companies might say, hey, we don't want to, you know, the agents for conducting the interviews. We just want this scheduling agent that will help us really offset this scheduling tasks of our recruiters so they can, you know, focus on the more meaningful activities or they can go full-on and they can say, hey, yeah, we want to go from the scheduling to screening to interviews to scheduling and everything in between. And this is very important to define because once we define this process, the second thing we do is that we need to define what we call success metrics. Okay, so we have the process we want to tackle with AI and what we want to achieve with this process. This is also one of the most important aspects because if you don't know where you're going, you're just going to be lost on the way, right? So we are defining these success metrics and these are revolving around three areas, usually. Now, but it's straightforward. First one is what an impact the AI will have on the recruiters workflow. In other words, how much time it can save in admin tasks for the recruiter so they can really focus on the more impactful work? Second is usually how the AI can impact the candidate experience because there is lots of talk about AI interviews. For example, some people like it, some people hate it and it's really, really, really hot topic these days on the market. So we can define, okay, so what effect the AI will have on the candidate experience? And third one is usually about quality, right? Do we hire faster? Do we have better pipeline? Do we hire better talent and so on? So we usually defining metrics along those lines and once we have them, then we can launch the kind of first pilot, let's say, and the pilot we recommend running for six to 12 months because first three months, the things are selling down, right? Because the AI brings really fundamental shift to the recruiters live. So first three months are onboarding, training, getting them up to speed, how to utilize the system and the next three months are about harvesting value. So the recruiters, the organization already know how to use the system, how to deploy it in different stages of the hiring process and can start harvesting the first value of the system. So this is what we usually do, but the really the key lesson for me is that it's not about AI, it's about people, it's about humans and our psyche and our willingness to change and our willingness to be open for innovation. 100%, I think that is kind of the real sort of crux of all of this. But how is this changing the recruitment process? Because obviously you talked about the different stages that people could use. If someone's using all of those agents, how does it change the way that the process works? That's a good question and it doesn't have to change the process at all. There might be the same old process you usually run, just be run automatically by AI agents, but I think or we think that it's a kind of bad thinking because with AI, we should start thinking about AI first process. And yes, currently, I'll say 90% of organizations are not ready for AI first process and they are implementing our AI to their current process or old process, but the same steps like first round, second round, third round, you know, final interview, offer and so on. But we get from the first yearly adopters of our fairly new functionality or feature that I personally really like because we call it instant job interviews, which basically shortens up the hiring process to just one touch point with the candidate, which is amazing. What it does is that you can either create a QR code or get a link from talent pilot, you can put it to your career page, job description, or job posting, whenever you are kind of getting candidates and this link will lead to the AI interview with our AI agent. And the, you know, candidate doesn't need to go through the boring form all the time, you know, filling in the name, surname, email address, LinkedIn profile, cover letter, CV resume and so on. Just click on this button, the AI will, you know, introduce the opportunity. The candidate can ask any question regarding the company culture, values, strategy, the team, the job itself and so on. The AI will answer and then the AI will ask a few questions relevant to the requirements for the positions relevant to the skills that I required to be successful in the role. And at the end, if the candidate is satisfied with these answers, by the way, the candidate receives a real-time feedback from the AI. You are very strong in this area. You might get to work on this area slightly more and so on. And if the candidate is as happy with the outcomes, then he just say, yeah, I want to apply for this position and the AI will manage the rest. So you can imagine how this speeds up the whole process because after just one short touch point with the candidate, you have all the information you need as a regular throw hiring manager. The candidate is happy because he applied. So obviously he's looking for the job and he thinks that he is a good fit. And you can then directly schedule a meeting with the hiring manager. And the process that, if the candidate would fill out the form on the career page, they would go either way through some type of interview, either pre-screening with the recruiter or screening. We just would really make it in one step and say time for everyone involved. I suppose the candidate experience part of that. You've touched on that quite a lot. I just want to sort of pull that out a little bit because you also mentioned that there's a whole narrative about people not liking AI interviews and things like that. And I think when you look at it, there's a very easy media story that people don't like AI interviews and all that sort of stuff. But actually, I'm speaking to lots and lots of organizations doing this kind of thing. And the reaction tends to always be actually quite positive. So I just don't think that we're doing a very good PR job around the candidate experience benefits. So here's your opportunity. What are the benefits from the candidate from having AI managers this part of the process? Sure. That's a great topic to cover. We are currently working with one of the biggest banks in Europe. And since the biggest banks, they are rich. So they have lots of money. And they were willing to spend the money on very, very diligent UX testing of the AI interviews. So we've invited people from all sorts of demographic areas, like senior people, senior lawyers, senior AI engineers to really salespeople across genders, age groups, and everything. So we have told a few, say it was around 30 to 240 people in this UX testing. And we were literally sitting in the dark room behind the mirror, like an FBI, and looking how the people react to the AI interview. And they didn't know that they are doing the interview. They just knew they are going for a hiring interview. So it was very interesting to see. First of all, how usually people, once you tell them it's AI interview, they expect a chatbot. They don't expect this natural human voice like AI that reacts to your voice. You can interrupt it. You can ask it, whatever you can go offline and it brings you back to the topics like really human words. So they expect a chatbot that they will be chatting with something. But after the AI interview, what actually surprised me in a good way was that the AI raises emotions in people. Because when I was sitting in the dark room looking at the respondents talking to the AI, I saw by myself, when the AI will praise them for something, for some experience, they will say, "Yeah, I did this and that. I achieved that." I said, "Hey, that's a good, it's an amazing experience you got there." And you can see how they start smiling and how they shine up all of a sudden. So that was very interesting. But after the study we did, the findings was very clear. First, one of the most important aspects is people can take the interview whenever they want. If it's 2 a.m., if it's 10 p.m., you know, whenever they want. And you know it yourself, usually your best people, your best workers don't have time to take interviews during work day because they want to deliver the best value. So we know in our own data, we see it in our own data that the best people are taking the interviews at 9 p.m. 10 p.m. After they came back from work, put the kids to the bed, and then they have time for themselves to prepare for the interview and take it. So this is one thing. And second thing is that pretty much everyone from the study group tell us that they believe much more in this AI interview from the fairness perspective that they are beating, being treated the same way as any other candidate being asked the same question as every other candidate and half a fair chance of getting the job. So I think these are very, very important ones. Yeah. And I think and as you already said, everyone is getting instant feedback, which is the thing that I think people, you know, people want communication, people want people want feedback. And I think this is, you know, this is a route to do that at scale, but human, human recruiters just couldn't scale up to do. So I just think that's a kind of a huge thing as well. And in terms of the value that it's driving for employers, what is that? Is it kind of efficiency? Is it speed? Is it quality? Is it where are your clients getting the kind of most value at the moment? Honestly, all of the above, we can really see and there are already starting to pop up scientific papers from large-scale studies implementing AI in the recruitment process that there is lower attrition of the new hires when the AI is involved. There is the process is faster. In our specific case, we can reduce the time to fail by on average 50%, which is amazing. If you have time to fail 50 days with us, you can have 25 days, which is a huge time savings. And obviously, the most meaningful to me at least is that the recruiters, the people in the organization actually don't have to do the heavy lifting, the admin work, but can really focus on creating meaningful relationships with the candidates, creating meaningful journeys for the candidates, because it's all about employee branding. Even if you reject some candidates, you want them feel that they've been cared good after, and that they were, even though rejected, they will still be promoters of your company. Absolutely, and I suppose that brings us on to exploring that sort of role for recruiters, because you make some great, great points there. But, you know, there's a lot of stuff that we've talked about that recruiters would see as their role. Finding candidates, making decisions about who to take through the front of the process. So, how do you kind of see that role evolving? You talked about humans in the loop. What is it that recruiters are going to be doing? We see with our own clients how the role of a recruiter is evolving significantly. We can see how it evolves from, let's be honest, mostly junior admin heavy role to one of the most impactful and strategic roles in the entire organization, because with town pilot, the recruiter needs to start thinking strategically and conceptually of what type of outcome they want to achieve with the hiring process, which is something very, very different from what they do now, because they are currently like living day to day. I'll wait for what CVS will arrive today on my desk. I will screen them. I will do the, you know, prescreening calls and so on. But with town pilot, they need to think what type of person I need to hire in order to, you know, achieve these goals, and how I will tell it to the system. How I will explain to the system the goals the system needs to achieve with it. So, now we call it actually a super recruiter, because a super recruiter will need to think about this type of outcome in play, then manage the AI system, and the AI system then, you know, does it autonomously, automatically. But what's more important is that we believe that this type of super recruiter is best suited to have this, you know, role-defining the outcomes, because the recruiter is hiring across the organization, across levels, across departments. So, you know, hiring managers, they need, they know what they need from their, you know, isolated perspective. But the record is since hiring through the, you know, department level and, you know, individual levels, they know what the organization needs. So, they need, you know, combine these together and define the outcomes for the AI to then achieve it. Finally, everything that we've been talking about so far is here right now, you know, you're talking from your experience of implementing this for a number of large employers. Where are we going next? Where might we be in sort of a couple of years time in terms of the sort of the capability of AI and these, this type of technology? Where's it taking us? Well, nobody knows. It's a good, good philosophy question. Honestly, I think that the job market itself will look like nothing today, because we are in this, I call it transitory stage, is currently, you know, the first touch point between an organization and a candidate is, or might be, you know, AI interview. That might be the kind of initial point where the person talks to AI. And I'm not saying it's good, because people like to talk to people. And this is where I think it goes. And we are starting to see first companies that are building agents that are representing candidates on the job market. So you, as a math, you will have your own AI agent that will know everything about you, your personality, values, preferences, you know, skills and competencies, and so on. And then the other side of the marketplace, the companies, they already have these AI agents like talent pilot. And then the first touch point between the company and the candidate will actually be without people involved. It will be just two agents talking to each other, hey, I'm agent from representing this bank or representing talent pilot. Hey, I'm agent representing math. Okay, let's talk to each other. Yeah, I have this opportunity for math. It looks like a good fit. So let me schedule a meeting directly to the recordless calendar with math. And both the reporter and to you as a candidate will just receive the, you know, scheduling emails to your calendars. And the second touch point will be person to person. I share that and agree with you a million percent. I think really that I think that's where we're, where we're going. Where are we going? Because I think that just makes real cruising better for everyone. Tom, thank you so much for talking to me. Thank you. My thanks to Tom. You can follow this podcast on Apple podcasts on Spotify or wherever you get your podcasts. You can search all the past episodes at recruitingfeature.com. On that site, you can also subscribe to our weekly newsletter Recruiting Feature 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. My show.