SPEAKER_02 0:00

One of the watershed moments was this synthetic avatar application. It was a person. They were speaking naturally. They were sitting in our digital interview and they scored in the 99th percentile. We were interested to speak to the candidate. Unfortunately, their phone number was like one, two, three, four, five, six, seven, eight, nine, ten. You perhaps see individuals from a lower socioeconomic background, staff, perhaps a little bit more timid at the beginning. And by the end of the week, you could see a real change in their confidence. We're a company that are fully embracing AI. We're investing 500 million pounds in digital. If that doesn't sound like that's right for you, then perhaps we're not the right place for you to come and work for.

SPEAKER_00 0:36

AI in assessment, allow it or ban it?

SPEAKER_02 0:39

Allow it.

SPEAKER_00 0:40

Skills or qualifications. Skills. And finish this. The CV will be in five years.

SPEAKER_01 0:49

Still here.

SPEAKER_00 0:54

Hello and welcome to the Talent Download Podcast. My name's Darren Lancaster. I'm the CEO of Amberjack. Thank you ever so much for either watching us today or listening to the podcast. I've hope you enjoyed previous episodes. Today I'm really excited to introduce you to Rob Davis from Grant Thornton. Rob's been on a really interesting journey with Grant Thornton. If anybody's seen anything within the press, they will see their approach to early careers is really, really interesting and the way in which they're increasing head count is quite significant within the area. But one of the things we want to talk about today is the AI assessment skills model, which has been created alongside Grant Thornton via Amberjack. And we think it's a good topic to discuss to everybody listening in. But before we do that, Rob, always good to give everyone sort of an introduction, understand a little bit more about your background and how you kind of got ended up at Grant Thornton.

SPEAKER_02 1:50

Great. Thank you, Darren. It's a pleasure to be here as well. So I started in recruitment when I came out of university agency side, as most as most people do, I think, in recruitment these days. Did agency for a good couple of years. I I don't know whether it was foolishly, but started my own recruitment agency when I was like 24, seemed like a good idea at the time. You realize looking back, it was way too soon to do something like that. But I don't regret it. You you learn loads of things from from doing something like that. And uh it certainly felt brave at the time. But um uh after that, I I moved in-house and um and started a Grant Thornton. That was 10 and a half years ago. Uh I guess there's something in a TV that kind of been at one place for for so long looks a little bit peculiar uh in this day and age, but four different jobs over the course of 10 years there. So it's always kept it very fresh. Um and and today I'm um in the head of recruitment uh role that oversees recruitment across training all the way through to directors.

SPEAKER_00 2:53

Isn't that strange that you say um that you've been in a role for 10 years like constantly, but that seems a bit strange today. That's so different, kind of slightly older, but it was such a strange concept, kind of because when I entered into the workplace, it was all about you know, longevity, you stay in a job for kind of for a long time, but that kind of has been now kind of accepted that that ability to fluctuate. But I think the rapid change within organizations that you kind of work within has kind of it's just meant you've done lots of different things, kind of being there and not quite clear your promotion throughout the organization.

SPEAKER_02 3:25

Yeah, and if you you find yourself in a place that you you know match your matches your values and kind of has goals that align with yours, then um providing yeah you know you have the right attitudes, you're curious, and all of those sorts of things, you can you can find your way into different things quite quite nicely, I think.

SPEAKER_00 3:41

Good. And so from your um perspective, how we we're here to um sort of unpack some of the things that are kind of happening from AI. Um, but how's that um changed candidate behaviour um towards kind of AI tools um and in in the mainstream? What have you you kind of seen?

SPEAKER_02 4:01

Yeah, uh I think the um the biggest thing that we're the two biggest things I think that we're seeing from a from an AI piece with candidates is volume, first of all, in terms of just the the ease at which they can now apply for multiple multiple positions. So we've definitely seen an increase in uh in applications, which puts a lot of stress on on the kind of operating model that we that we would have had that we would have had in place previously. Um the the other piece that it does as well is um the the quality of the application flows further down the recruitment funnel. Um when I think about experienced hire, uh people are using AI to match their CVs a lot more closely with job descriptions, therefore on paper they're looking a lot stronger and requiring a lot more sort of screening time from our team. In in our early careers uh world, we're seeing some amazing innovation from candidates where they're using avatars and putting lip syncing into it and content over the top of it and um almost having a completely synthetic application. So um you see lots of weird and wonderful uses of it and each impressive but equally causing challenges um that previous models weren't set up for.

SPEAKER_00 5:15

Yeah, I I think everybody, as you as you touched upon, everybody looks the same, don't they, in terms of that application process. And I've I've spoken to many organizations that that aren't as sophisticated enough in their processes, they've either sort of tried to do it manually in the in the past and try to deal with that application for and they just say to me, Darren, everybody's the same, you know, everybody's the same, that application. So that's the thing that you've got to kind of think uh to unpack uh around it. But what um what do you think kind of the risks that kind of are presented from from AI in the process?

SPEAKER_02 5:51

Um uh I'll start I'll start with the risks so that we can talk about the opportunities hopefully for a bit longer because that's always better. Um I think the risks is is entirely that you're not getting a genuine representation of either somebody's approach to the way that they work or their capability of of being able to do the job. Um so people can people people can get so far down the funnel, and really, really clever people can get a little bit further and and and get in through the door. Um, so that's probably the biggest risk in terms of getting people on board that aren't the highest quality candidate and aren't able to to kind of fulfill the role that you need to do. Um there are a whole host of other risks in terms of like we we try and kind of sweep things up in background checks and things like that, so people aren't kind of um uh leaving out information that's important for us to know as a regulated business, for example. Um, but but generally we're our kind of North Star is always finding the best talent and the best person for the role. Um and if we're not doing that because AI is masculine, that's puts a big risk on our business because we're a we're a time business, we sell our time. Um, so it has a direct impact to the to the top line.

SPEAKER_00 7:06

So Grant Thornton um helped with the shaping of the AI skills assessment tool. Um and I I think it was you that kind of came to the team and and initially sort of had a discussion regarding it. Talk us through that story a little bit of kind of the thought process that went into it.

SPEAKER_02 7:26

Yeah, um, it would be completely remiss if I didn't mention one of my colleagues Louise in in this conversation. Um, but uh Louise and I had done quite a bit of work probably post-2023 on kind of defending our processes against AI. We kind of had the the boom of ChatGPT and you know, universities and employers all all getting in in a little bit of a panic around kind of how we could allow individuals to use this in our processes. So um I think we'd spent probably a year, a year and a half or so focusing on defending our processes so that we were getting authentic candidates through. Um, and then we were talking about well, I think one of the watershed moments was this synthetic avatar application that we got, and it was a it was a person, looked like a person. They were speaking uh naturally.

SPEAKER_00 8:21

Oh, so that came then, that kind of came before the audience.

SPEAKER_02 8:24

Yeah, there was content. Yeah, they were sitting in our digital interview, um, and they scored in the 99th percentile. So that was a really like wow, wake up moment that kind of this is moving really, really quickly. Um, we were interested to speak to the candidate. I was gonna say, did you hire them? That's kind of the kind of question. The first thing that came into my mind was let's speak to this candidate because surely we can find something for them. Um, unfortunately, their phone number was like one, two, three, four, five, six, seven, eight, nine, ten. So they they obviously didn't want to be um uh called. But that was a piece where we were kind of like, wow, there's there's some real capability and a real skill in terms of people being able to use this. So we kind of flipped our thinking to right, we've got the kind of defense sorted, sorted, and we can, we're really comfortable and confident knowing that we're gonna get authentic people applying who want to work for us, um, who have the right kind of skill sets and approaches to work. And then it was like, well, AI is going to be a really important skill set going forward and people's ability to use it. So, how can we kind of walk the walk almost and say, like, yes, we're an AI company, we like using AI, we encourage it, but then we don't actually assess for it at all. Um, and we're defend uh defending against it. So we kind of thought about well, our case study in our early careers space is a really high predictor of quality in in uh performance in role. It always has been, and it's almost like a simulation. So, how can we create as as lifelike a situation that somebody, a candidate, can go through that it feels like they're at work for us? So they will have a real a real world problem that we have to deal with, and then they will have the tools to answer that problem. And the big the big piece there was one of the tools that we all had at that point, they did not have. So they didn't have a large, long language, large language model to help them with some of the kind of work that we would be doing. So we thought, okay, we'll we'll go and we'll go and speak to our current provider and we'll ask them to just add this tool into the into the assessment, and it wasn't there, and and they didn't have the capability to do so. And then we explored it a little bit further, and it wasn't anywhere in the market at that time. Um, and that's what led us to a conversation with you guys. We spoke to a couple of um businesses, and I think what really stood out for you for you was you were going to build the whole thing from scratch, from the ground up, and not superimpose it onto something else that was already built. And um, you know, it's been a really great 12 months of kind of scoping that out and building it, and we're right on the edge of having a product that's ready to go to market now. So um that's kind of the the journey here.

SPEAKER_00 11:00

Yeah, no, I'm I'm delighted you came to us, I guess, in the first place to kind of kind of talk it through because it's almost moved on. I think it was all almost 12 months ago, almost 12 months now, or something like that. The kind of the discussion uh almost happened. And all of the research and the studies that that that come out now is is all talking about how do you basically utilize a large language model alongside you, work with you to help you in the workplace. It's no longer an organization that's trying to defend it. If you're defending it, don't worry, the employees worked a way around it and has kind of got their iPad or whatever next to them, alongside actually providing that. So you were kind of forward thinking ultimately, you know, as an organization to think that through.

SPEAKER_02 11:43

Yeah, and we kind of thought the power in it was in simplicity of it, it's it's just replicating what somebody's going to be doing if they were in a chair as an employee as a candidate. Um, and and it we'll we'll be able to iterate it in terms of the way that we've built it and without getting too nerdy and uh in into the into the detail of it, that the way that it's set up is is is in a safe environment. So everybody kind of has that level playing field across it as well, which was obviously really important when you're when you're building these things.

SPEAKER_00 12:11

Yeah, I think it's important that the model, you know, it's you you're you've built it with a your large language model that's sitting on it. It's not like someone it's is based upon someone who can spend whatever 60 pounds a month on Claude or something like that and get a larger model um base within it. It's a very fair playing playing ground that someone's actually utilising the same tools around it. Um but so the the challenge that you were you were trying to solve, you were you're in a position for quite a period of time of defending against AI, like many organizations were uh at that time, to then kind of get into a position to say we need to change this, we need to mix this up overall as an assessment because it wasn't addressing what you basically needed from it. Yeah, yeah, it literally came from. I I I think it was just a really important point that you you got to basically as an organization to actually see that and see that through. Um, did you get inside influence through the kind of design process? You kind of once you started the process, what did you learn?

SPEAKER_02 13:14

Yeah, I think it I think like every the great thing was everybody had like ideas on on what this could look like uh along the same principle, I think. So um we uh I guess part of the beauty of the partnership that we have is is we have been hands-on and we've been able to kind of talk through ideas. Uh, you know, you have a great team that we're working with that are kind of back with loads of research, and the RD team, as you mentioned previously, have kind of been involved in that as well, and and obviously working with the the the the people to to build the product. It's um for somebody that doesn't, I would say that I don't speak tech, I'm just curious. Um, so sometimes these conversations go down a track and you kind of lots of things are going to be.

SPEAKER_00 13:55

You're learning, you're learning during them, or you just you slowly googling or chat GPT and yeah, yeah.

SPEAKER_02 14:00

But I would say the team have been really good at kind of um talking about it in in layman's terms and and and just the way they talk about the large language, large language model, so that it's fair for everybody. It's the phrasing that they've been used is there's a harness around it so that it kind of there's a complete control, there's a control so that it's geared towards the assessment, and you can't be using it for kind of loads of other things that would just kind of swamp the amount of energy that's needed for it and things like that, and kind of cause fairness issues. So loads of involvement, and um, it's been really fun kind of working. Usually you work on these types of projects in-house with your teams, and internally it's been really nice working with with another company, kind of sharing ideas and different approaches about it as well.

SPEAKER_00 14:42

And what's your thoughts of the general marketplace? You talk to lots of organizations out in the market at the moment. What do you what's your views on the way in which they're assessing talent?

SPEAKER_02 14:51

Um, I think there's a real want to to kind of assess AI capability. I think because of the way it's changed so quickly, I think that there's almost been a bit of a wait in terms of waiting to see, well, well, let's see if it flattens out at all so that we can kind of get a good idea of what actually we need to assess. And whilst it probably doesn't show any signs of flattening out, it seems like it's still going even quicker. I think broadly people are are understanding now it's it's there's an ability to prompt, like there is there are skill levels in someone's ability to prompt. Really importantly, it's can you can you assess critically what the output is and can you make sure that it kind of is geared to what you do? So, you know, that they're two of the key things that I think are coming out right now that people want to assess. And it's just there's a a lack of lack of technology platforms or whatever, and until recently, obviously, that to um to do that. So I think people will catch up, and I think we are in a constant game of catch-up at the moment, whether you like it or not, it's it's it's an interesting place to be, but you're always kind of waiting to see what the next iteration is or what what it can do next. I think it'll probably move on to prompt efficiency and things like that. One of the things that we we've launched recently at Granthaunt is is a clawed license for all of our all of our UK um employees. And at the moment it's kind of carte blanche, it's like go and experiment, go and go and try with it. Um, and and we will kind of be educated, I suppose, in due course around kind of tokens and how much it takes to kind of do certain things on different models. And I can see there kind of being a commercial element to it going forward where you'll need to have some sort of efficiencies around your prompting um and not just have unlimited prompts to be able to do something.

SPEAKER_00 16:44

Yeah, no, it's an interesting point around tokening, and that's probably another potential podcast or something more to talk about regarding it. But if you the more sophisticated the models are getting, the more obviously that's draining your own usage uh around it. If you kind of move to Sony and kind of clawed out ultimately, you're draining your usage and and how do you control that ultimately across your workforce? I think is one of the challenges that people are gonna have to be thinking about. I mean, even myself personally, I kind of I utilize I think three different models for different kind of work output regarding it just to be efficient ultimately.

SPEAKER_02 17:20

To it's really funny. Our um our chief digital officer who's who's not long joined us talks about the different models. So we talk about Claude, but the different models in Claude, he he almost talks about them like a garage, a car garage. So you wouldn't take a Ferrari to go and do your weekly shop as in in kind of in that sort of mentality. So I think the the most powerful way is to kind of I was gonna say try and make it relatable. We're not all driving around in Ferraris, but um people can kind of relate to that idea and kind of understand when they should be using certain models. And um, I I just think that'll be important the next step.

SPEAKER_00 17:56

And how do you think um we should think or employees should think about um candid performance when basically utilising the models? Is that something you you're thinking differently about in Grant Thornton?

SPEAKER_02 18:08

Yeah, we're we're um we're obviously interested in seeing what their prompts are and and seeing what they're what they're taking from those prompts and putting into kind of an output. But I think um one of the things that obviously we we've talked about and um with the project that we're working with you guys is kind of hearing from them about what they're meant, what their thought process is around what they're doing. So uh the ability of of being able to record them talking and saying, this is why I did this and this is why I took this. I think just seeing that mindset and that approach around it is is as helpful as seeing how they're actually using the AI as well. That's given us really valuable information and and kind of a a really good layer of scoring on top of just being able to see their prompting and what their outputs are.

SPEAKER_00 18:53

And what do you think the the human aspect of the skills that need to come into the role? Uh what do you think is required?

SPEAKER_02 19:00

Um, I mean, in in terms of this, it seems like we've had various conversations over the the last few years. And um I think we probably base a lot of our human skills around uh particularly in the trainee space on what the human skills the trainees need at that time. Um, I think we're probably moving to a world where we're hiring trainees who are going to be our future leaders. So we're probably looking for an element of leadership quality in there. I mean, as I said, we're we're we we sell time as part of our business. So there's an element around relationships and things. Even when we've kind of talked about how we're evolving our own recruitment team, it's like how can we evolve our recruitment team based around those human pieces around relationship building? Um, because if we start building our team around kind of doing administrative tasks, then then the team will crumble quite quickly because AI will be able to do that. So um we're kind of we're still in that in those conversations where we're kind of moving from what are some of the kind of human skills that trainees traditionally needed in a in a pyramid model, as most consultancy and and professional services firms are, to what are some of those human skills that we need to be kind of on the look at now for our future leaders?

SPEAKER_00 20:18

It's a lot about um the five C's that you can kind of think through. It's about the communication of an individual, it's the curiosity, uh, it's the creativity and then that of an individual, which becomes kind of the key part of what you need to be looking for in skills going forward. Uh and I was it was interesting, I was talking to uh Springpot, who obviously do quite a lot of work with working with universities, colleges, schools, uh, an introduction regarding that, and that's what they're trying to educate through kind of to the school systems is basically these very human uh intelligence is basically what's required for the future, and that's what organizations are ultimately going to be looking for.

SPEAKER_02 21:06

Yeah, and and curiousness is one of our one of our three values as well. And um, whenever we bring new cohorts of of talent in, um, you know, we'll have various board members talking about talking at events, and all of them will mention curious, be curious, ask questions. If you if you if you don't know the answer, you must ask, because you don't always get a playbook in for for roles these days. It's not just uh this is a manual, how to do the job, it's uh it's you should have a base knowledge, but being curious around will build that knowledge and and make you a better person. I think it's probably a common trait in most of our high performers in in our business, I'd say.

SPEAKER_00 21:46

Yeah, and and uh going back a little bit more towards the product um and and thinking through those kind of human aspects between it, um how do you kind of distinguish between Whether a candidate is just a really good prompter to the other kind of skills or assessment level, which would traditionally, I guess we would consider the things that we'd look at, maybe future potential or different models.

SPEAKER_02 22:15

Yeah, I mean, so so we we look at it across a I guess a kind of span. Um so uh our assessment, you know, our case study is one component of our assessment. And I think what we possibly might have done in the past, not overly done, but we might have tried to assess all of the um the attributes or skill sets at each point of our assessment, um, each different point of the assessment. I think where we're getting more comfortable now is that there are certain parts of the assessment that are better um suited for assessing certain attributes that we're looking for and other parts of the process where, for example, a case study will assess will be heavily based on assessing someone's capability around AI usage, but we also assess all a few of our other pieces in that in that approach. And and um but a big thing of it is the critical thinking piece, like the AI skill and the prompting thing is is is the kind of the fundamental piece. But all of all of the work that we've done around internally launching Claude at the moment has been you own the input and you own the output to make it really, really obvious that don't just fling a prompt into the AI and pull it out and just put it into something else that you that that it's intended for. It still requires that level of judgment. And I suppose, particularly for our audit audit practice, which a majority of our trainees come into, that's what clients are paying for, they're paying for the judgment. Um, and it's it's making sure that they're able to apply that, and and that's certainly the one of the other things that that's heavily involved in the case study.

SPEAKER_00 23:52

I think there was there was some great sort of press releases, I think it was around about Easter, Easter time, in terms of Grant Thornton and their their thought process around early careers, and I think that's how they're increasing by 30% uh across the board um for that workforce, which was fabulous to see because I I'm quite passionate about some of the doom and gloom that we kind of sometimes get within the press about AI's eating up all their jobs, there's gonna be no future for early careers. I really you can directly correlate that back to how that's affecting young people, and I just think it's the wrong, and it's not actually data-led ultimately from what's actually the reality, certainly within my clients and the data that I see uh across Amberjacks, and we we work with you know a considerable number of clients across the board, and it just doesn't add up the numbers. Yeah, yeah, sometimes there can be a decline because of um what's happening in the economy, and that that can often be of course that's the case, and that's the case on all jobs, uh, ultimately. But it it was good to see that positive T from Grant Thornton. But of course, that was strategically led because your views on what the future kind of your business needs to be.

SPEAKER_02 24:58

Yeah, and I think on that point, you know, you you'll always see uh trainee volumes, particularly where you're hiring a high volume of trainees fluctuate over a kind of a period of time. I still think we're probably still in professional services, still feeling a ripple effect of of COVID, where after COVID there was a massive boom and a and a huge uh hiring in consultancy and things like that. And now we're seeing it it kind of contract. And that's not just at the trainee level, but you know, we're seeing some of our competitors um uh really kind of trying to pull the strings in on kind of all grades across those areas. So that there's a ripple effect still still being fed, uh still being kind of um felt from from that point, and then you layer in AI and the kind of uh things that I think it's probably just an easy thing to kind of talk about at the moment. Everybody's talking about it, and uh, but but there are definitely other things that affect kind of our decision making around how many trainees we're gonna hire every single year.

SPEAKER_00 25:56

And talking about those individuals that are coming through, what do what do you think the skills that they need to think about um going forward? What should they be developing if I'm I'm currently a graduate enroll? Um, I've got one of my sons is in there at the moment um and and doing that, and I'm trying to trying to guide him down the right path in the things that he should be thinking about. But what is it from your side?

SPEAKER_02 26:18

I I think um, you know, you you talked about the five C's and and kind of obviously the the book that's um out there from the LinkedIn guys at the moment. That you don't need to go a f a long way away from those. I think communication is important. I mean, there's the stereotypes around the generation at the moment uh that probably pretty unfair a lot of the time. We see some really great candidates come through our process. Communication's really important. Um, just being able to kind of hold a room. We have group exercises, being able to interact with with others is is one way that we kind of test that. But that's a really important one. The curious curious is is there and and embedded in our values. So we think that's really important. We've um we've probably talked a lot more about courage and and confidence, I suppose, in the in the last few years at Granthawns, where we're kind of on this journey of taking on investment, being one of the first more one of the first largest, larger kind of um professional services firm to take um private equity investment and the kind of courage that that required to kind of go into uncharted waters. Um, so a lot of those things kind of resonate with some of those human that the more human-led uh attributes that um that that require and and that kind of follows through to kind of what you want to see in your in your in your candidates and your trainees coming through.

SPEAKER_00 27:39

Yeah, I think the the practical experiences um that you you can see from uh the future workforce is important. Um I've mentioned it on on on the podcast um before, but where my son's uh first sort of thought about doing a graduate program and it had in it in it had this placement year in it that said, you know, you're going off into the workplace for a year. I kind of when I first looked at it, I was kind of like, oh, that's like four years, kind of that feels like it's taken a long time. But what's actually happened is the reality of that process that's come through is they've got this kind of real life work experience of working and actually they're in finance and accountancy. Um, and through that experience and that life, and and you know, being there, they're in being told to go in the office five days a week during that um placement year, but building up some of those personal human skills around that office interaction has done wonders for for them, anyhow, in terms of really understanding what the workplace was, because that was what the workplace was when I started around it, but it's it's those human characteristics which are so important.

SPEAKER_02 28:46

It's it's a great point. I I mean, I'm relatively new to early the early career space, as I mentioned, kind of experience higher was where I cut my teeth, and um I'm probably three or four years into early careers, and one of the things that I was really interested in when I first came in was kind of our conversion program. So we have a 12-month placement year where uh where candidates will come in from university, work with us for 12 months, and then hopefully they enjoy it and go back to uni for a year and and come back to us. Um, we did internships across the summer for for four or six weeks and kind of tried to convert those. And that to me seemed like a really great way to hire people at that sort of, I'm generalizing now, most of our trainees are kind of that that younger age. Um, partly because I did I didn't know what I wanted to be when I was 18, 19, 20, 21. And I'm not sure I could have coped in the environment that we put our our kind of trainees in at the moment. But um, so we look to extend our conversion program as much as possible. So we have multiple ways that people can come in and spend more time with us over that longer period, one to see if it's for them, but also for us to assess kind of what their growth might be like. And um, we relaunched uh our work experience program uh a year or so ago. Uh, we call it Ignite, and we bring in um six-form students, they come in for a week. And um whilst we give them lots of practical work to do, you can really see the growth of individuals over a week as opposed to a 50-minute assessment. Um, the really stark thing was as well, you could see individuals coming from private schools were on day one, suited and booted, ready to go. The finished article, um, when it came to kind of how they were communicating, what they knew about us, all of those things. You'd perhaps see individuals from a lower socio-economic background, staff perhaps a little bit more timid at the beginning. And by the end of the week, you could see a real um change in their confidence and how they were kind of um interacting and kind of the work that they were doing. And it and it was it was really great to see that because then we were able to not just look at some metrics on an assessment scorecard, but we could really see the growth over a week, and people could say, if they can improve that much in four days, how much could they improve in three years or five years? So um we we do try and focus a little bit more on those conversion programs. Uh we, you know, there's loads of research out there that you know uh that their people's brains aren't fully developed until they're 23, 24, 25, 26. Uh we're assessing them at 18 for a job that's a corporate job, and they're advising, you know, companies on what they should be doing with their finance departments. So um we felt that was the the appropriate way to assess people across a little bit of a longer period to see the potential as opposed to just what they're answering on a answering on an isolated part of our assessment process.

SPEAKER_00 31:52

And you you quite clearly very involved with the tool, you know, really helped us as a business. Do you see the tool just operating within the early career space or do you see it across all its sort of experience hire, specialist hire? Yeah, yeah.

SPEAKER_02 32:07

I mean uh the the the the way that we've got our our particular assessment set up, it it's just a business problem that um anyone could deal with pretty much in in our organization and it and it and it could be used by by probably probably anyone, to be honest. Um uh we we've just done our first set of value uh validation. So I've just got off a call before coming to you today where we just did our our first set. Um, and that was really interesting to see how individuals went because there was a nervousness because when we started building it, it's like, well, there's an advantage here because they're gonna have a large language model. We have to make it harder. There are certain things that we're gonna have to do, like increase the amount of data that they have to absorb to the point where it's impossible for them to read all of it. Uh, we wanted to make it contentious so that AI didn't naturally give them an answer. It kind of was a little bit reluctant to um to do that. So there are things that we've done to make it more challenging, and we were always mindful that we hadn't tipped it, tipped it too far. Um, but the early set of validation shows that we hopefully we've got it, we've got it right and um and and that's promising.

SPEAKER_00 33:21

So um yeah, I I I I being part of it, you know, equally having to go through it myself and and be part, I I think we have. I I think it it generally is uh around it. So it's uh it's just really making sure, you know, the guardrails and and those types of things, which you know, quite clearly we want to put through it, through its pace is easy are in place. And um at Amberjack, you know, we've we've been utilising AI to help in the screening of video screens um for for for quite a period of time now, and we'd learn lessons from it, and it's really about making sure you've got your quality assurance based around it, and as I say, those guardrails in place to to test. Um, and I think too many organizations don't get that right ultimately.

SPEAKER_02 34:04

Um yeah, definitely. I was asked recently kind of what what would what would your advice be to somebody that kind of was embarking on kind of setting up the AI in their assessments? And um, like you, we we just didn't start using AI in our assessments two or three years ago. We've been using it for since 2018 in digital interviews. And um the the I wouldn't say lessons learned, the things that we've been really, really cautious of with doing that is adverse impact and making sure the guardrails are there so it's fair. Um, I think without that bedrock of experience and knowledge around it, it would be quite easy to make mistakes. Um, if uh if you were just embarking on this for the first time and were just looking for something impressive to to mark a candidate and um to get them to use AI. So uh, like you guys, and and again, that was a another advantage of partnering with you, is is is that added experience as well in in terms of um ensuring that it's a robust product.

SPEAKER_00 35:03

And when when you're talking to that a little bit more, when you're embracing the AI um or use of AI, do you think you lose anything from a fairness or authenticity kind of basis on it? Or, you know, obviously it's been sort of front of mind, I guess, ultimately.

SPEAKER_02 35:20

If you're gonna use AI in your assessment, do it for a reason. I mean, we're an we're a company that are fully embracing AI, we're investing 500 million pounds in digital. Um, so it's it's embedded in things, everything that we do already, and it will only get more. So, so if AI is not your thing as a business, then like you need to question whether or not you you really, really want to kind of um not necessarily assess it, but I think what we're able to do is be quite be quite firm in our messages to say, like, we use AI, it'll be part of your everyday. If that doesn't sound like that's right for you, then perhaps we're not the right place for you uh to come and work for. So we can kind of really uh leverage that um that piece, and and I think that's important to do because otherwise you're just gonna get people that won't kind of um commit to it or or whatever.

SPEAKER_00 36:19

Yeah. And and thinking and looking ahead for the future, what it's a hard thing to do today, and and have that kind of crystal ball in many ways, um, and thinking about where the businesses will be. But from an assessment perspective, where do you think that will be over the next sort of few years? I might I I dread say you could use say like five years to people, but I mean it's I just almost find that's too far. Is it's almost like you know, one year, two years, yeah, ultimately.

SPEAKER_02 36:44

Um yeah, uh yeah, we we tend to we tend to we tend to we'd be very good at working in years in a professional services firm, it's a partnership, everything is kind of yearly focused. Yeah, okay. I think in in five years, where I'd love it to be is um you'd like to be able to go to the four corners and be super accessible for um candidates to come to you and be able to deal with volume in a really confident way, um, without being worried by it. Because at the moment it's uh in volume hiring, the more applications you get, the more money you're gonna spend on recruitment to kind of get to get it down a funnel. So um I'd I'd love to get to a place where we can where we're not um nervous at least um about volume of applications and really confident that we can deal with them in terms of the skills that we'll we'll be assessing, like who know who knows that there's there's always a podcast with someone that's kind of looking into the future and saying something that you think is 20 years away, and then they say it's four years away, and and things like that. So um I think that's probably what we're we're probably aiming for at the moment.

SPEAKER_00 37:53

Good. And um what's one thing that organizations should stop doing when they're assessing talent, do you think?

SPEAKER_02 38:02

Um probably go back to my other point, my my point before, where um understand what you're looking for and don't try and assess for everything. Um like I said, if if if you're not using AI in your business, don't you don't necessarily need to assess for it. If if some of the some of the things that are kind of topical at the moment aren't some things that kind of individuals in your in your uh organization need, then then then don't assess for them. It it's been really, really deliberate. We spend a lot of time thinking about what are the really core values, core skills, behaviors that um that impact performance on individuals as a firm uh our firm, and we design our assessments around those things. We we try not to kind of do a a broad brush approach. We try to be really deliberate in terms of what assessing what is successful, first of all, knowing that about yourselves, and then going out and really building a process and an assessment process that that brings all of that out.

SPEAKER_00 39:07

What do you think firms should start doing? What should they start thinking about?

SPEAKER_02 39:11

Um oh gosh. Um I I think experiment uh with it with kind of all the tools that we have available, but be very aware of kind of um uh of how it can be misused. I think some people will kind of get Chat GPT of Claude or Gemini on their phone and kind of there's carte blanche on that. It's a personal use, you can kind of use it for anything. Um, you'd probably have to be pretty, pretty careful about how you're using an assessment in a professional capacity. So I'd I'd I'd say always make sure that you're very aware of the risks and what our obligations are around you know, candidate data and how how we're treating candidates fairly. Um don't just rush into it and just think we're gonna get people to to kind of use AI without kind of the right guardrails in there.

SPEAKER_00 40:06

Yeah, good. Um always like to finish the podcast with some quick fire and questions. So one, two words are uh are allowed. Um let's try with one anyhow, but uh uh forgive me for looking at my iPad as I fire through these. So I'll start with coffee or tea? Tea. Morning or afternoon interviews?

SPEAKER_02 40:29

Uh morning.

SPEAKER_00 40:30

AI in assessment, allow it or ban it?

SPEAKER_02 40:33

Allow it.

SPEAKER_00 40:34

Uh skills or qualifications? Skills, potential or experience?

SPEAKER_02 40:39

Potential.

SPEAKER_00 40:40

A video screen or in person? In person, structured or unstructured interviews.

SPEAKER_02 40:45

Uh I quite like unstructured interviews.

SPEAKER_00 40:47

Yeah, that's recruiting. Uh gut or data?

SPEAKER_02 40:52

Oh. Um that's a difficult one. Uh I will I would go with gut. I like data though.

SPEAKER_00 40:59

Okay. Uh finish this sentence for me.

SPEAKER_02 41:02

Employees should measure more more about their values.

SPEAKER_00 41:10

Uh, one thing every assessment should test, but usually doesn't.

SPEAKER_02 41:15

Um I will say courage.

SPEAKER_00 41:22

And finish this. Uh the CV will be in five years still here. Okay, I agree. Well, thanks ever so much for talking to me today, Rob. Um I hope everybody enjoyed listening to the podcast or actually watching it as well. Uh, we're super proud uh of the work that we've been doing with Grant Thornton on the skills-based AI assessment. Uh, if you're interested in the product at all, please directly contact us um via our website, and we'd love to talk to you more about it. Uh, it's something that's been created um with a lot of support from Grant Thornton in terms of the way in which you see the future of work, which we fully align uh on. So we thank you for that. Um, and we really do hope it can benefit um many organizations in the future, and then we further sort of develop and advance it from there. So, thanks again for listening in. Uh, look forward to you joining us on the next podcast very soon. Thank you.