Fullständig transkription av 2 MILLION Applications: How Deel uses AI in Talent Acquisition with Alan Price

it was a massive expansion. So building up a lot of EMEA hiring from London to Germany, France, Italy, Spain, know, then moving into Nairobi and Egypt and all of these markets. So was a real, yeah, eclectic mix. George Kelly: Amazing. sounds like some some really good experience, and obviously, you know, some exposure with some of the ⁓ the best known brands. ⁓ And ⁓ having rec well, I say recently ⁓ joined Deal, but I think you celebrated your three year anniversary not too long ago. Is that right? Amazing. Alan Price: Yeah, three years now. So I joined April 2020. George Kelly: and hiring over 200 new employees every month. Alan Price: 2023? Yeah, 2023. it's gone incredibly fast, So when I joined, I think we had like 250 million around that mark, think, like annual reoccurring revenue. Now it's well over a billion. And yeah, the company was about 2000 staff. And now we're like seven, seven and a half thousand, I think. George Kelly: so if you're building a team leading recruitment, Alan Price: Yeah. George Kelly: It's incredible growth, honestly. And and if we take a few steps back to when you joined years ⁓ what was it about the vision that you saw with within the company and the the leadership team that made you commit to this challenge? or trying to figure out where AI fits into your hiring playbook, this one's for you. I'm excited for this episode. Let's dive in. Alan Price: Yeah, so number one, you can't not pay attention to growth, especially when you're in recruiting. If you're growing as an organization and your revenue's growing, that's really cool. The second attraction was, it was profitable. the last 10 years, Uber, we weren't profitable. At Walt, who then got sold to DoorDash, we weren't profitable. electric scooters and bikes. We weren't profitable and it was really refreshing to be part of a startup scale up where it's actually profitable and kind of makes money. So that was cool. then from a connection point I really resonated with At Uber we were trying to scale and grow into a number of different markets from a recruiting perspective we had a couple of challenges. Number one there was a lot of friction regulatory wise and we would have protests, would have strikes, don't know if anyone remembers like in London for instance, taxi strikes across the city is a big thing that was replicated across Spain, France, Germany, know Italy all of these kind of major markets so trying to convince someone to join an organization which was under severe regulatory scrutiny was a challenge. Like, hey, come join us, but aren't you illegal is the narrative. ⁓ so that was quite challenging. And then number two, ⁓ from a HR perspective was hiring people without an entity and then trying to convince people to leave their consulting gig to move and create their own entity, doing a... ⁓ thing ⁓ and convincing someone in Germany to you know leave well-paid job the security and then come and join this organization you know we can't even give you kind of employment terms, Until establish the entity, all of these things. So, Dill was really solving a problem and I that, you if I had this 10 years ago, it would be absolutely game changing. So, that's what really resonated for me. By the time we were only doing, we were only Contractor and Employer of Record. Those were the two products, nothing else. George Kelly: if I sort of look back on because I remember first hearing Deal towards the back end of the COVID pandemic coming out of it, where the work from home narrative started to become a more popular way of way of operating, right? Alan Price: Yeah, I think like people often misconstrue like, like, deal as a poster child for kind of remote work, right? We're the largest global, fully distributed remote company that's out there of today. A lot of companies have that hybrid model and all of these things and like do allow for remote work, but consistently 100 % of all of our workforce is remote. child. It's not just, our clients want remote work, therefore they're going to hire an employee of record. That's like some but not all of the cases, right? Cases, you know, I'm a US business. I want to expand into Europe. I want to build an office in London, but I don't the right kind of, you know, entity setups or, like infrastructure. So solving those types of problems is not necessarily like you have to be working on a beach somewhere to be remote work. know, it's really not quite that. George Kelly: Mm. Hmm. we've seen Deal evolve massively over the past few years and the the breadth of services that you guys are is incredible. and this not only goes out to the actual employees and the the workers also businesses and the partnerships that you're developing globally. it's fantastic to see you're managing roughly one and a Half million workers worldwide. I'm sure those numbers are increasing day by day, but that's pretty impressive numbers to be managing. Alan Price: bigger when ⁓ you look at the payroll, right? So when you sign a global deal to run payroll for 20,000 person organisation, those numbers tend to quite quickly. George Kelly: Yeah. I imagine so. Yeah. And so from a talent perspective, how many applications are you managing on an annual basis? 'Cause it's like it must be close to like two million now, ⁓ Alan Price: Yeah, so last year we did 1.3 million applications and hired 2,900 staff. So the decimal placing goes really far back. So people would say like, top 1 % talent is like 0.00, maybe there's another 0, 2, of application through to higher rate, So 1.3 million 2025 in 2026 Q1, we had half a million applications. So that is going to trend to about 2 million this year. We're not going to be doing as much hiring towards the back end of the year. So it'll probably come down a bit, but current run rate is about 2 million. George Kelly: that's unbelievable. And if we look at most companies, one of the key issues that we face, particularly in the GCC, is actually application volume and how

you navigate through all of the noise. And I think we'll touch on this in moment around how AI could potentially help with that and give a fairer opportunity for applicants. But equally, most companies would panic at that volume. So I'm really intrigued to know how ⁓ you've enabled Deal to actually move faster and ride that wave. And how are you not drowning in in that volume? What does the structure look like to filter through that noise? Alan Price: Yeah, so it's a multifaceted problem by the way, and we deploy AI in a number of different areas and we deploy it not to make decisions, we deploy it to give us better signals to be able to find great talent within, know, when I say needle in a haystack, it really kind of, yeah, it really is, right? So. For us, let's start with CV review, you all of your job requirements, all of your competencies. You weave that into your... AI bots, to look at CV review firstly, to up candidates, right. Then got your database search. Database search is really powerful and we get lot of hires from our existing database of millions of candidates where they might have for a account executive role previously, but their background is SDR. We wouldn't hire a new account executive who hadn't done that role before. So ⁓ again, if we're to a sales development Representative for instance we could look through our database see matching against that they may have just applied for the wrong job or Might just been closed when they opened right? And then you've been got outbound as well. So we do a range of market mapping research Automated outbound so when we calibrate the box and make sure that they're kind of working well Then we're able identify people you can automate messaging all of these things make sure that we're reaching out to the right folks So on the piece, ⁓ is a heavy deployment. I'm sure we'll come onto that in a moment. ⁓ Then you've also got, heavy deployment of AI during the interviews. So ⁓ we've played with and done different things. I don't want to call it an interview, but I want call it like a filtering system, So... I'll give you an example. So we were doing a world record attempt in January for the largest global hiring event ever recorded. we did AI and two things there. Number one is we did. an interview with an AI agent, not as an interview, I don't want to call it an interview, like it's more of a screen, right? To able to match up the CV and match up a few questions, To calibrate because what we sometimes find with like sales development representatives, the signal of the CV isn't as good as the signal of like the person. So you want to get to know them better. And we definitely do that for our interviewing process, but. Who should we be interviewing at the start? the CV, combining a short AI exercise to give us a better signal as to who to invite to meet the recruiting team at recruiter screen and then follow through into the interviewing process. ⁓ then number two is when we actually go through those ⁓ interviews, then deploy Metaview, which is an AI tool for note taking, but we've got a lot of statistical analysis and insights and reporting and all of these things. I'm sure we'll come onto the techniques of that later. ⁓ to be able to make sure that are we assessing the skills that we said were really true? What was the quality of question types? What was the talk time? What was a number of different aspects around what makes a quality interview? And making sure that we refine that over time and enable our hiring teams to continue to be better at interviewing. Because if we can deploy AI in a much more meaningful way for interviewers to be better calibrated, George Kelly: Hm. Alan Price: that we'll make better hiring decisions over time, not trying to use AI to evaluate the candidate. We're actually trying to enable the interviewer. ⁓ George Kelly: I that's a really interesting point you've made there, actually, because one one thing that, we've noticed is that the use of is enhancing ⁓ non recruiters, i.e. the the hiring managers ⁓ actually recruit better and interview better. So what would you say you've noticed in own hiring managers and what have you been able to analyse through their interviews and that analysis of how they're conducting their interviews Alan Price: give you one example. So in our customer success area, right? A lot of the question types around development, because in customer success, for instance, ultimately is to grow accounts, So you might couple of products and how do you position our existing product, ⁓ existing or new products to make sure that they can provide value to our customers, Because we all see fragmented system across the people ⁓ tech and we want to enable better support ⁓ our clients. a connected system because what we tend to find is a lot of specialist systems get plugged together. They all serve a different purpose, but there's not a of a uniformity and they don't always speak to each other, which is kind of really important when you start to deploy AI insights and all of these things across your entire people stack, ⁓ necessarily like this one tool for this and this one tool for that. So we see ⁓ much more. consolidation of the tech stack. So that's number one, right? Now, number two, when we looked at strategic client development, what we found in our interviews was lot of the interviews was experience led. ⁓ a past example. Tell me a time when you've grown an account, right? Every single person has experience of, tell me a time when I've grown an account. Doesn't mean they're good. That question is not a good leading indicator ⁓ potential performance. So when we looked at and evaluated that everyone was given a great answer to, ⁓ there was this one time that I did a great thing for this one client, right, and we grew it from this. George Kelly: Mm-hmm. Alan Price: But that doesn't mean that there's behaviors and consistency to it. So what we actually saw was it wasn't predicating outcome,

i.e. folks coming in, and it wasn't a good indicator for us. understood that ⁓ in 94 % of they were either asking a historic question or they weren't using probing or two or three questions afterwards ⁓ on that one topic to evaluate someone's behaviors and how they would solve a future we just changed the question to a future problem. And with that data, it was really insightful for us to be able to measure that. the data previously was we would a scorecard. There would be a rubric on there and it would say, evaluation of client development. And someone would tick yes or strong yes with an example of in the prior role they did this, right? So they evaluate and said, yes, this was a really good competency, but the question was wrong. George Kelly: Yeah. Yeah. It's always down to the questioning, isn't it? And unless you have the analytics, it's very difficult to measure. So it Alan Price: Yeah, I I always say, I always say recruiting is a black box. You don't know like before or the tools that we have. you would go into an interview or the candidate would go into an interview with the hiring team and you would just get a scorecard at the end, but you wouldn't know what actually happened during that interview. either, like previously I've sat in interviews because it was not calibrated, I've sat in interviews myself, in years gone by and like with the hiring, that's so inefficient. ⁓ the tools that have available at our disposal is game-changing, right, to uncover the black box of what goes on during an interview and then be able to provide that enablement and support to make sure that we're, like TA, the guardian of the process. We're not the experts, we're the guardians. George Kelly: Mm-hmm. Hmm. I like that. That's a that's a really nice perspective, And so if we look at the key traits that you're looking for at deal when hiring across the board, because I'm sure that there's a couple of sort of fundamentals that you track ⁓ across all your employees, right? What what do you typically look for and what do you filter for? Are there sort of two or three signals that you and your team filter out, do you think? Alan Price: Yeah, so... ⁓ I mean, simply put, you know, I'm very simple, So I try and be as simple as possible with a hiring, So three main categories we look at, One is domain expertise. So depending on the role, if you're a software engineer, it's about coding, right? If you're a salesperson, it's about sales ability and acumen, right? If you're customer support, et cetera, et cetera, So you've your domain expertise there. Number two is you've got behaviors and cultural values, So you test for, through our Elevate program, we test or cultural values and behaviors, right? Along the entire interview process as well, by the way, these aren't just like rigid. Number you've got AI fluency. So AI, a of are wanting to use AI in a much more meaningful way. So we have to test for AI fluency today to make sure that these folks ⁓ are... embracing the

use of AI because as an organization, we're embracing. So therefore we want our workforce to be embracing of AI versus being scared of AI for instance. So those are the three main areas and then you can dive deeper, What are the competencies that are assigned to each different area? then we take assessments. ⁓ form part of our hiring ⁓ And then from that, then, when come into, ⁓ know, a worker, when they've just joined us, how does it measure against the existing scorecard, right? So one or two months. We're still debating on the exact, like we started with one month kind of signal and then three months. We're debating if it's like two months is better, right? But getting an early signal ⁓ hey, ⁓ is how we evaluated the candidate when we first made the offer decision. ⁓ Are these competencies, and you go over a little bit deeper, are these competencies the same as what your forecast is? the recruiter is always a forecast, and when you're here, ⁓ it turns into an actual. ⁓ So then you can some of those gaps of like, hey, I thought this person was going to be exceptional. ⁓ And it's not quite, still a good hire, right? But it wasn't the that we thought. And likewise, if we someone was like, yeah, meets expectations, ⁓ needs a little bit hand-holding, but ultimately, you know, we'll be fine over the long term. And if they come in and they're much better than we thought, we can evaluate those gaps of like the forecast to actual, and refine our interviewing processes and techniques to get signals and get better hiring. And that isn't measuring candidate, that's measuring our interviewing signals and how we actually get signals from our interviewing process. George Kelly: Yeah, I like that a lot. And and one thing you you mentioned just then was around the measuring of AI fluency, right? And I we're seeing this more and more often now with with companies that are hiring for non-technical positions are actually looking for this capability and enablement of AI within non-technical functions. And ⁓ think that that is so crucial for the listeners to take note of because as we're evolving and as AI becoming more and more adopted and it's becoming more and more important. And what we're see starting to see as well is non-technical ⁓ staff actually building solutions for their own pain points internally. And ⁓ I'm sure that same methodology is being adopted Deal. So I'm really intrigued to know ⁓ you know what does that look like? Are you actually looking for individuals who are willing and interested, and curious to actually build ⁓ with AI internally ⁓ opposed to relying on on their engineering or tech team to build the solutions for them. Is that something that you are working on Deal Alan Price: Yeah, so I there's a couple of different aspects. So we test across, we're obviously at testing with our engineering team, That's already there, using AI for a while, right? for our non-technical roles. So just explain a bit of context. So single ⁓ person at Deal,

⁓ have access to Claude, ⁓ ⁓ chat functionality, through to copilot, through to Claude code. ⁓ it is the expectation of every single person ⁓ use AI. And we measure who's using it. We credit to everyone. They use credits. And then if they want more credits, ⁓ then give them more credits. I'll give you an example. In TA, we had meeting across my team two weeks ago. And the meeting was, was every single person on my team and it was, what are the problems that we see today? ⁓ We don't want to go into every single problem, but I want every single person to pick one problem that they see and I want you to go build something in either Lovable or Claude present your findings back in two weeks. Now I don't mind if five people independently work on the same problem, it doesn't matter. The goal isn't about the output. The goal is build one thing, you're more likely to build the second thing. ⁓ likely to build the third thing and get better over time. If you don't build the first thing, you'll never get to the second thing. So got the pitch ⁓ by TA, ⁓ mirroring the by Deal, which George Kelly: Yeah. Alan Price: ⁓ that we launched for any startup founders to go access to, I think it was like 15 million or something ⁓ of for their startups. So the pitch by TA was whatever the ideas we'll them. We'll turn them into a product, right? ⁓ Working with the ATS team, with our internal teams, on like building a product based all these areas. So next week, I'm going into an evaluation. I'm looking at I think it's now 29 or 30 different projects that folks have concentrated on or worked independently on some of the things just Phenomenal and this is really good in terms of ⁓ So that's mirrored across a number of our teams so yeah, we want folks that are gonna buy into that culture and George Kelly: Yeah. Alan Price: buy into that kind of methodology, right? And I think for me, it comes down to like a rubric, And when you look at that rubric, you've got like the basics, like a lot of people just think AI is about chat. And it's like, if you're just ⁓ looking chat to optimize an email or something, or like form an on something and you know, then... ⁓ fine but that's not really where we see the power of AI, right? George Kelly: Surface level, right? Alan Price: Yeah, surface level is like, and then when you say, ⁓ how much you use AI? Yeah, I love AI. Great. What do you do? Well, how do you use it? What was the business you're trying to solve during an interview? ⁓ it sometimes it just comes back as basic and you're like, okay, well, you what are the four different levels? We think about it in four different levels, right? What are the four different levels to actually like scope where people are? Are people using just chat functionality as kind of level one, Level two is then, ⁓ you know, being able to co-pilot certain things all the way up to level four, is like augmentation and running AI independently on a variety of different work streams that have broad scope and have a much more meaningful impact to the business if you can kind of augment certain things. George Kelly: Mm. I think it that's such an it's so so interesting. I mean, one thing that I've realized on that point is that I think sometimes people tend to sort of over overcomplicate or overthink AI and how to actually adopt it, how to implement it into their day to day. When the reality is, is that the limitations just your mind and and how creatively you're willing to think and then how you then go about instructing the AI to build, right? So the limitations are endless. If you can think it, you can make it. And that's just the world that we're in now. Alan Price: the same as going to the gym, ⁓ You haven't been to the gym for a few months, ⁓ out of bed at six o'clock in the morning to go to the gym before work is really bloody important, right? And then you go and like... ⁓ The first time to the gym is really hard. You're really like, oh, this is a struggle. Then the second time to the gym is like, oh, this is still a bit of a struggle, but like, I'm getting there. And then as soon, and it just gets easier and easier and easier. And then all of a sudden you like start seeing the results and the incremental gains. You're not going to be like an expert on AI in day one. So just go and make a few mistakes. Go make a few mistakes. It doesn't really matter. And if you just have that culture of like, I wanted to build something, I just don't know how, Like, some of my team said, we don't know how to use it. I was like, great. Learn to learn, right? George Kelly: Mm. Alan Price: ⁓ There's a school in Amsterdam called ⁓ Kodam, ⁓ the founder of TomTom, right? And all about peer-based learning. There's no teacher. It's just you get into a class, there's an objective, go build an app. ⁓ there's teacher. And then... the students have to figure it out collaboratively together. So they'll go do Google searches, someone might have some expertise, someone doesn't, they all like train each other on how to actually, objective is clear, go build an app, right? Or go build a website or go build whatever it is. And it's based on like the level 42 principle ⁓ of different the methodology is like to learn. So when someone says to me, don't know how to do something, it's like, go ask Claude. ⁓ George Kelly: Mm-hmm. Alan Price: Go ask your colleagues. If I'm just giving a top-down thing of this is what I expect you to do, that's not peer learning. So you can go and build Claude. You can have two different windows open. One is like Claude Chat, and the other is Claude Code. It's like, hey, I'm trying to build this. I'm working with Claude Code. Give me hints and tips. This is what I'm trying to do. Ask me questions you're fully clear and you're not making assumptions. ⁓ George Kelly: Right. Alan Price: this I'm trying to do. It will then ask you some questions and then you say, hey, how do I give a good prompt to Claude in order to start building me this thing that I want

to build? Then you can use them together and then you'll start to learn more. You can save your chats in artifacts. There's various different plugins. There's a ton you can do, but you just need to get started. George Kelly: Yeah, yeah. hundred percent, Alan. And something that that we've been doing heavily we're building out our own internal skills that are super specific and super detailed. But as you say, like you just get started. ⁓ And once you started, Alan, like I'm sure you've seen this as well, but you kind of get hooked a little bit. I find myself, you know, up for till ⁓ early hours of the morning just you know, building, building, building. ⁓ And one thing I wanna put out to the to the listeners most a lot of the time when I speak to talent leaders, often get a little bit overwhelmed with that's out there, right? It's it's all a little bit confusing for for somebody who maybe is non-technical. So, you've already on but where where should they start? You know, where where do you recommend them getting started? Alan Price: is a problem you're trying to solve? Pick one problem you're trying to solve. You can sometimes pick a topic that's so broad and it covers compensation, it HR operations, covers HR, it covers employment and legal, it covers like, it's a multifaceted problem. Don't start there, right? Start with one problem you're trying to solve and then think about architecture. Number one is Claude gonna do this for you? Is a product or software you're gonna have to go and buy, is that kind of getting better for you? And then you think about solving one thing first, and then think, Okay, if I'm going build this, how does it scale? That's the ultimate goal, right? So does it scale? But firstly, is there a is AI intelligent enough? What the capability of ⁓ the thing that you're trying to build? What platform you doing? Right? you know, is it chat GPT? Is it Claude? Is it what are trying to And ⁓ get started on something, right? And then from that, it will just make your life easier. I've ⁓ got where ⁓ we've Metaview, which is our video interviewing tool, George Kelly: Yeah. Alan Price: Then I've connected it through to an MCP ⁓ and then I just ask a number of different questions every single week and I get a daily digest of what's going on during our interviews. And it's just a reading thing that I take a coffee in the morning, I just have a little scroll and I've just asked it, me the top five things ⁓ are going on in the interview, And then how does it compare to last week's top five things? ⁓ Give a breakdown. And it's really interesting. interesting to just be able to see and compare and contrast and then to be able to kind chat and ask questions to it it's just insightful. It's just automated. ⁓ just get that it's coming through on a daily and weekly basis. George Kelly: Yeah. there's so many benefits, right? And I think what it enables most is this clearer view of processes, but also the inefficiencies. And like you say, it's just start really and those incremental steps and just keep testing, keep trying. And what I found to be really useful as well. I don't know if you do this, but I often just braindump everything I'm thinking, because we can get overwhelmed with I just speak to Claude. And I just brained up and and it I then bank that over across a week and it compounds, compounds, compounds. And then during that week I can then extract pieces of information from that bank and say, hey, let's let's develop this a little bit further because I think it's got some we it's got legs, you know? Alan Price: Yeah, exactly. it comes back to culture, And for me, where I found success is like speed, right? So one of our values of deal is like deal speed. shipping is more important than the actual product that you're doing. you do ⁓ you can iterate and you'll get to a better result faster ⁓ try and develop something for a month. think it's great, haven't tested it, you finally come to launch and then there's all of these gaps that like all of these are unintended consequences and all of these things that you didn't have the foresight to think about and like it's not about having the foresight because nothing is ever launched perfectly. Every product launch, everything, there's always things going on behind the scenes, if you compare it to, I'm just gonna launch something within two days and then I'm gather the feedback and iterate it. Instead of taking you a month, it's gonna take you like 10 days in order to get that's better version than what you would have taken a month for. and it's the old adage like, whenever start a company, it's, if you're proud of your first product, it means you haven't shipped it fast enough. So, try that lens. ⁓ mean, don't know how universal that is, but that's a widely viewed analysis of if you proud of your first product, it means you haven't shipped it fast enough. Try that lens ⁓ to your deployment of ⁓ If you're proud of first thing, you haven't been able to iterate it and make it better over time, it probably wasn't good in the first place George Kelly: Mm. 100%. It's just about shipping quickly. It doesn't have to be 100% perfect. Get it out, get it into the hands of the users, use it yourself, validate yourself, and then you know exactly where the problems are. So you can iterate quicker, improve quicker, and so on and so forth. 100% it makes sense. ⁓ I'm intrigued now, because we've we've touched on AI a little bit, but we mentioned earlier around the Guinness World record that you set a deal, which is Quite an interesting route that you took. So I'm intrigued as to how AI was used within that and what was the actual kind of purpose behind it? What were you trying to prove within that project? Alan Price: It's a great question. So I'll tell you the origin of how it evolved because I think it comes back to shipping first. So we've got our headcount plans coming through where we wanted to front load a load of hiring for sales in the early

part of the year. And it's really important for us to hit hiring goals because every month that someone's not in seat, it affects quotes for entertainment, So when we evaluated, looked at, well, hey, there's like 250 roles that we need to hire for, and they're all pretty much front loaded Q1. We got the hiring plan through, in December. like, ⁓ How are we going to do this? We just did the analysis of, well, hey, if we need to interview 12, 13 people for a role, multiply that by how many outreach and applicants we need, how many recruiter screens, how many interviews, conversion rates, scheduling, all of these things. We're going to need a bigger boat and we're going to have to change strategy. ⁓ And number was that we didn't have ⁓ a lot the interviewers like capacity wise in order for them to even conduct interviews. right especially for a hiring manager if you're growing a team like 30-40 people there's a ton of interviews you gotta do we said well hey let's just do ⁓ ⁓ call with and just get them excited ⁓ because a lot people ⁓ just think that is an employer of record slash remote ⁓ facilitation ⁓ company we to change ⁓ the brief and the narrative. So we were just gonna, for some of our markets, so Japan was the first market. like, hey, we've got some roles in Japan. ⁓ Sometimes we struggle global business ⁓ going into employer branding, all of these things, right? ⁓ we're like, let's just do a call in Japan. And then that was just from a recruiter saying, hey, I want to do a call in Japan, like, are we good? I'm like, sure, yeah, problem. And then we're hey, what about Australia? ⁓ Yeah, we kind of to, ⁓ we're going, ⁓ okay, well, what about GCC? What about EMEA? What about ⁓ LATAM? What about us? This is big, like, ⁓ So I was like... ⁓ I wonder what's biggest event that's ever been done? I just, I just, just clawed at it, right? And said, ⁓ what's the global hiring event that's ever been done? Like, just curious. ⁓ And it like, ⁓ there's no And I'm like, I there's no record. ⁓ There must a record. ⁓ What recruiting world records there? Couldn't find a single one. ⁓ George Kelly: I'm gonna set it. Alan Price: Okay, cool. And I was like, how do we a record? How do we get a world record? It said, okay, well, this is what you need to do. So I got one of my ⁓ recruiters ⁓ to find the person in UK who ⁓ Guinness World Record, get me telephone number. we phoned up the person, got the telephone number, old school, phoned them up. pitch them as they're going into a meeting say hey we want to do a world record with you this is what we're trying to do and he's like the hell okay send me an email I'll put you in touch with someone. So we were like, okay, great. And then we were relentless. ⁓ got calls, ⁓ start December, right? ⁓ Then two weeks, it was it was a thing, they had to go through... like the legal teams and make sure that if you're setting up a new event there's a lot of structure and like what are you actually measuring definition of the actual attempts so we went through that took a couple of weeks and then we're like by mid-december we're like all systems go and we're like okay well when are we doing this event it's like 14th of january and like ⁓ hmm ⁓ we've no so we started to use AI bots ⁓ to ⁓ mass campaigns we started a linkedin campaign ⁓ and honestly George Kelly: Ha. Alan Price: Honestly, because the world record wasn't created, we agreed on a number of 1,000 people turning up for the event. And we're like, okay, we'll over 1,000 people. And then we're like, okay, well, we expect 20 % response rate from outreach. So let's see if we can get to like 3,000 just to give us a little bit of buffer. We ended up getting 30,000 applications. for weeks, for sales roles. We got 16,000 registrations of which it ⁓ eight a half thousand people turned up the actual call. ⁓ Obviously from time zone perspective, there were a number of other candidates that were like, hey, time zone didn't kind of work, but I'm interested kind of thing. And that was the objective. It wasn't the objective to get as many people onto the call. The objective was to create awareness, so even if you didn't attend the call, it didn't really matter because we'd still put you through the programe And then so we had eight and a half thousand, I think, people attended the call. The world record stipulated that you had to have at least 30 minutes of people like duration for 30 minutes on the call. So it was like 6900 and something of people that were over the 30 minute and Then from that use to do the We use AI database search and inviting people into the into the process. We then use the AI for a CV score Then we used AI for the signal that I talked about earlier, Going and doing a screen with an AI agent, it took about 15 minutes. We asked three or four questions just to give us better signals as to their suitability for different roles. We then got those signals and then selected people to come into our interviewing process in which they met with a recruiter, the hiring teams, hiring managers, department heads, elevate interviews, and then presented it off. So we hit all of our hiring numbers. that was ultimately the goal. We hit all of hiring numbers and achieved the world record that like this didn't exist in recruiting before. Right, so I think it was just that was that just shows you the speed of execution. We wouldn't have been able to do it without AI like but would have had to be months and months and months of planning. We're just like, let's go. George Kelly: Yeah. I love that. It's pure ⁓ and it was a way of drawing a huge amount of attention in a very ⁓ short sort of compacted of time ⁓ and and you achieve goals, which was the main outcome. So that's that's fantastic. Alan Price: We got so much, got a bit of heat for it as well, right? ⁓ George Kelly: But it's also a good heat, right? It kind of like a little bit controversial. Alan Price: I didn't see the heat of the sun as necessarily good, but like

we got to think of this is so inhumane, it's a cattle market, you you've got 30,000 applications for all of these jobs, you there was, you know, you publicly promoted that there were 7,000 people, you know, attending the event, all of these things, this is like cattle market, right, recruiting, this is so inhumane. I would attest standing, not even standing by this, like I would attest that this is the most humane thing that you can possibly do when you're dealing with that volume and let me that down for you. So we gave opportunity to people that would otherwise discount it through those conversion rates of having ideally like 10, 11, 12 recruits accrues to one hire. you can't go and and interview everyone, right? Everyone fair access to great paying jobs. And also recognize that the CV doesn't always correlate to performance, especially in an entry level role. So being to disconnect the at the first evaluation point to see who you invite into an interview to get signals, we actually found that a lot of the people who performed really well during the AI screen were coming from ⁓ different backgrounds. on their CV. So actually the inclusion part of being able to give fair opportunity to everyone in a ⁓ demotocratic kind of way. George Kelly: Yeah. Alan Price: actually enabled us to spot different talents. We had accountants that were coming and saying, hey, I'm done with accounting, I wanna be in sales. We had baristas, we had a ton of people, and they actually showed their competency during the interview and rethink the valuability of the CV and experience. For that, Then you're able to give a wider selection and also you're able to then use AI. So you're not using it for administrative tasks. You're using to actually connect with people. Because my belief is if you can take time, give recruiters back time, they can speak to more people, give more fair opportunity. They can speak to people during the middle and end of process through how was your interviews. if you reallocate that time, you're able to put the candidate first during the process. and provide a much more humane experience through those processes. And in terms of the cattle market comment is like, we all see applications rising. If we're going to pass over and say it is not a free for all at the moment, it is wildly inaccurate. It is. George Kelly: Yeah. It's unbelievable. The volume of applications that we're receiving is crazy and it's super, super hard to manage. But obviously, you've worked out how to do that. And with the power of AI, it's just making it a little bit easier. But equally, without actually discriminating against candidates, you know, as you ⁓ correctly mentioned, you know, it's giving a fairer chance and more equal approach to the entire Alan Price: I'll give you this, our average to right? all of our roles, it's 30 days, right? And we that from when the candidate applies. So historically, you'd a seven day ⁓ SLA of like reviewing CVs. We're down two. ⁓ George Kelly: Application base. Alan Price: Then we speak to candidates that includes a scheduling piece of like getting a recruiter screen in the diary. Then it's team interviews and of the assessment that we put through. ⁓ then ⁓ it takes roughly about a week for your offer negotiation and like offer approvals, contracts, all of these things. ⁓ that 30 days is inclusive ⁓ when they sign. So. at our volumes, keeping it to two days half a million applications, every single quarter, getting that time to hire, taking all of the noise out, ⁓ and then our net promoter score ⁓ is 50, right, as a net promoter score. So we're investing time back to candidate experience. And we think, and even when we look at the disconnect between people that get hired and they complete a candidate and promoter score, they're obviously like more promoters than detractors, right? And then we need to look at the early part of the process. you have a lot more variance to like the detractor piece. But then we those data, we can look at those insights, we can connect one of the top main themes into candidate experience per area that then creates all of our problem areas. So then we can then start to address like, for instance, I'll you one more example. but ⁓ in our engineering area, ⁓ a interview. Sorry, would typically do a generic response, hey, thank you for your application after technical review, and the technical interview, and we didn't quite go how we wanted it to, please keep ⁓ it. So we're like, hey, candidates want that. So okay, well, let's just pick up the phone. And so because we were able to give more time back so they're not being drowned in applications, give them more time back, they can pick up the phone after a, if they want to, right? So we can say, unfortunately it didn't quite go as planned. Here's some feedbacks, You can structure feedback because you've got the recordings and you've got the insights, you've got the feedbacks, and you're able to give more insights to the candidate. We then rocketed from like plus 14 to plus 50, In the space of a quarter. George Kelly: Yeah. is truly incredible what it's enabling you to do. ⁓ And just especially at pace to also have time to hire at thirty days across the board, ⁓ like at the volume you're hiring at is is pretty pretty un you know, it's unfathomable Alan Price: a shorter right but an average across 300 eyes a month averages that ⁓ George Kelly: That's that's crazy. So look, I just just a couple of closing questions, Alan, because I know we're near in time, but how do you see talent acquisition and recruiting in three to five years time? And what role will humans actually serve in combination with AI? Alan Price: I mean it's big question for the last two minutes, right? I in general, I'll take a broader scope and I'll kind of go specifically on recruiting. The world is changing, right? The skills and evaluation of the teams, firstly, are changing, You need more AI fluency across customer success, customer

support, know, operational roles, not just technical roles. So that's kind of number one. So how you're testing needs to change. Number two is different technologies and products are currently being built. So in sales, for instance, there's a lot more software and AI that the teams can then use in order to be faster, better, or more costly, So the historic ratios of HeadCamp and SDR to AE ratio is going to change. So then it comes to workforce planning. just you have this role today doesn't necessarily mean that that is going to be that same role in three years time. Yet a lot of hiring managers are thinking about, like there's a backfill. I need this skill, this skill. So actually it starts from a workforce planning perspective. So imagine what future workforce needs to be, ⁓ have to play a seat at the table there in combination with your HRBPs and of that area, To really strategically think about workforce planning and where gaps are in the team. Then number two is how do you then operationalize that and make sure that you employ value proposition ⁓ of attraction, retention. We have all our OKRs based on four different categories. So our framework is called RISE. recruit an inspired workforce that stay and excel. And I think the stay and excel piece has often been taken out of recruiting. It's not my problem. Whereas actually you can't really argue of we recruit and inspire workforce that stay and excel. That's kind of mandate. So let's bloody test for it. And in terms of like tooling and all of these things. I think it's just going to be a lot more signals iterative points to improve, continually your quality of hire. I think that's where recruiting is going to be because... ⁓ It takes the edge of the five P's, right? If you have the right people building the right processes and building the right product, you align to the right purpose, then you get your profitability. But it all starts, if you don't have the right people, then you can't build the right processes, you can't build the right products, profitability and purpose if people aren't aligned to the right purpose. So it all starts with people. So, TA has been continuing. George Kelly: Mm. Alan Price: We all complain about like, I'll get a TA, get a seat at the table, all of these things. We have one. You have to sit down, right? Because the business want great people to build the right products, build the right processes and all of these things. So go sit down. George Kelly: Yeah. ⁓ Yeah, that is that's awesome. And actually that leads on quite nicely because look, the whole the whole reason why I'm having these conversations is because how you scale hiring is how you ultimately scale a company into ultimate profitability. So that is scaling with purpose is all about. And you've really nicely narrated exactly how you're you're doing that with deal. So look, Alan, it has been awesome. It's been an absolute pleasure. And thank you so much for sharing the story today. if you're if you're scaling a hiring function or building something, please do go and check out Alan on LinkedIn. I know you're always sharing some really valuable ⁓ insights. and ⁓ thank you so much. Alan Price: No worries and thank you very much once again for the invite.