Can AI Solve Recruitment’s Biggest Challenges? HMRC Says Yes!

00:03 – Lucy Adams (Host)
Welcome to HR Disrupted with me, lucy Adams. Each episode will explore innovative approaches for leaders and HR professionals and challenge the status quo with inspiring but practical people strategies. So if you’re looking for fresh ideas, tips and our take on the latest HR trends, subscribe wherever you get your podcasts. So there doesn’t seem to be any area of HR that isn’t impacted by AI, you know, whether it’s personalized learning experiences or employee benefits, better and more accessible employee survey tools, enhanced data and analytics, or just simply the automation of transactions. You know most HR teams are having to get to grips with the potential of AI, and today I’m joined by Andy Hedworth, who is the Deputy Director Talent Acquisition at HMRC, and Andy is trailblazing the use of AI across the recruitment function, not just within HMRC but also into some other parts of the wider civil service. So welcome Andy.

01:16 – Andy Headworth (Guest)
Hi Lucy, Nice to be here.

01:18 – Lucy Adams (Host)
Oh, it’s lovely to have you with us Now. You’re not from a traditional HR background, are you?

01:23 – Andy Headworth (Guest)
No, my background is purely recruitment. So you describe me in the civil service as a specialist as opposed to a you know, someone with a generic HR background. So it’s been recruitment right the way through, primarily the private sector. And then an old I guess an old customer of mine who became an HR director in HMRC. It was that call I had on a Friday afternoon. I’ve got a problem.

01:51 – Lucy Adams (Host)
One of those calls.

01:52 – Andy Headworth (Guest)
Yeah, and I went in for three months to come and help with some recruitment strategy and eight and a half years later I’m still there.

02:01 – Lucy Adams (Host)
Well, you’re obviously enjoying it and I think, from talking to you a bit earlier, my sense of one of the reasons why you’re probably enjoying it is because, actually, you’ve had the scope to do some amazing things. So let’s get into it. Let’s talk to us about the ways in which you’ve introduced AI into the hiring processes within HMRC.

02:26 – Andy Headworth (Guest)
Yeah, it started. It really did start two years ago. When you know, when ChatGPT landed. I’ve been an early adopter in recruitment all my career. So right back to when the internet came along, we were Googling for CVs in 2001 to social. You know, social landed in 2425.

02:44
And then you, you know, led that through and was very much part of of the social recruiting space, and so when a new technology comes along, it’s like straight away, you know, you’re looking for how you can use it, um, and and it, it’s unlimited. So straight away we got it, we were playing with it and and you go straight to sourcing and you play around with how you can find things and build strings etc. But primarily it started in about with my team about the Easter following ChatGPT, and we’ve got a group of us together and we just started playing with it. So we’d have an hour every week of all the millions of tools were coming out, literally, if you remember, at that time, probably 50, 60 a day and we were cherry picking them and playing with them. And then our little group grew to more and we had and we then started to think right, actually now we better get serious about this. We better think about. We’ve got a tool here that is going to be exponential. It’s going to sit across the whole hr and business space.

03:40
So we we very simply decided to chunk up our recruitment process to try and isolate areas we could use AI in at that time because it was still relatively well, still very new, to be fair and we sort of chunked our processes up into five or six generic process areas. We’d already been playing with sourcing. That was relatively easy to build strings, boolean strings, and help identify talent pools. That was relatively straightforward and that was very much independently led as a consultant or as a recruiter.

04:15
But we identified sort of two pain points. And the two pain points that we wanted to solve to start with, which causes a lot of pain in every organization, but certainly in HMRC was the hiring manager. How do we help a hiring manager do their job better, faster, quicker, save them time to do the things that they should be doing or they could do? And the second one was the world of talent insights. How do we use the tools and how do we build something to get effectively instant insights to be able to make a good decision and a hiring decision, um, or where do we go to hire people and do we move people around?

04:54 – Lucy Adams (Host)
so we picked on those two to start with as two problems um, interject there and just ask a couple of follow-up questions, because you’ve said a couple of things there that I think is worth exploring in a bit more detail before we go on to look at what you actually did. The first one was that you talked about playing. You know, experimenting with stuff and it sounds such a frivolous word, but it strikes me that there’s two approaches to this. You know you’re talking about 50, 60 products coming out every day.

05:23
I’ve seen some HR people just put their heads in the sand and just kind of, I’ll wait until the early adopters have got to grips with it and they tell me which are the best ones. But your approach is let’s have a play with them. Is that something that you would say is absolutely essential when it comes to a new area like this, that you’ve just got to experiment, try stuff out, play with it, knowing that most of it perhaps won’t work? Would you advocate that as a strategy if you’re thinking about introducing AI is just have a go A hundred percent, and that’s exactly what we did.

06:01 – Andy Headworth (Guest)
We were literally coming to, you know, we put an hour aside every week, which was effectively everyone came together and there was no real agenda. We we were trying to, as you say, find the tools that worked, find the, you know, the needle in the haystack with all of it, um and the and was basically full empowerment. It was come to play that. The tools are there, no one has got any rules, there is no right and wrong with it, with ai, and there still isn’t. Actually, there’s a lot of guidance yeah but there’s.

06:28
It’s still so new that people are doing some really cool things. Um, and so it was, you know, it basically empowered my team completely come along, we’re going to play. And they still have that empowerment now to to discover and go and explore the ai tools and how we can use them. So that’s never changed. Going back to your point on fun, that was the essential part. That’s what got people hooked, because we weren’t worrying about work at that stage.

06:54
All the tools that were coming out at that point were coming out, were making presentations, they were graphic related, they were tool related, they weren’t necessarily some of the hardcore things that are being built now. So we just had fun with it. We literally were just having some um, and then what happened was we? We ended up with a nice little group of people coming together and it was like, okay, we need to move this forward, we need to now stop looking at tools, because we had too many to look at. To be honest, they were everywhere.

07:27
So I said, right, okay, so let’s just, let’s have, let’s take, let’s put it up a notch, let’s see how, what you’ve learned in the last month, six weeks, eight weeks, and see what prompts you’ve done and see how you’re doing it. So let’s um, let’s have some fun, let’s just and it was a spontaneous in the one of our thursday lunchtime sessions and it was totally spontaneous, right, picture competition completely random, yeah, um, next week. Two, there was two pictures one recruiters in the wild and two a cat’s birthday. Absolutely no idea why I picked those two.

08:00 – Lucy Adams (Host)
What we did in the wild and a cat’s birthday. Yeah, okay, I’m hanging on, just yeah, keep going, andy so, but it was just, and even then, straight away.

08:12 – Andy Headworth (Guest)
Well, we I just literally said it and some of the guys and girls in the team were like, and pictures were popping up in the team’s chat, um and it it prompted a lot of interest and what happened was they went back to their desks and of course they were saying oh, we’ve just been doing this, I’ve just been creating this picture. And then someone else, the next week we doubled our people that had come along because they’re going oh, what’s this? This is curious.

08:35 – Lucy Adams (Host)
So then we so this wasn’t a Thursday lunchtime group just for your team. This was kind of anyone could come along and have a go.

08:43 – Andy Headworth (Guest)
To start with it was my team, but then there was no limit. So as people found out about it, people were joining from other teams, but they were all in the HR function. So, yeah, we didn’t limit it to our team, but at this stage we were still exploring ourselves, so we weren’t ready to sort of actively promote it.

08:59 – Lucy Adams (Host)
I think that’s a really important learning, isn’t it? That actually, to begin with, creating an environment where people can come along, get involved, have a play, experiment doesn’t have to work straight out of the box. It’s just about getting to grips with it and its capacity and its capabilities. But then the second point that you raised were as part of your process for introducing it was the pain points it was trying to solve.

09:26 – Andy Headworth (Guest)
Yeah.

09:27 – Lucy Adams (Host)
And it’s great to hear that, because I think quite often our starting point is what does the tech do, rather than what’s the problem we’re trying to solve. And so the pain points for you with the managers that you were working with. Did you go directly to them, or was this something that you were just aware of because you’ve been a recruiter for a while?

09:54 – Andy Headworth (Guest)
Very much the latter. We you know everyone in recruitment knows where the the challenges are is. If we can make a hiring manager’s life easy in any way, shape or form, we’re going to get a better working relationship specifically, were you trying to make easier?

10:04
so the problem they have with recruitment is they think this is it’s the volume, it the scale, it’s the. Oh, I’ve got to recruit again. Oh, where’s my job description? What have I got to do? How am I going to do it? Oh, interviews right, when are my interview questions? What do I ask? Oh, does that join up with a job description? Oh, and if I want to advertise on LinkedIn because my recruitment colleague is telling me to do so, what do I do?

10:24 – Lucy Adams (Host)
Yeah.

10:25 – Andy Headworth (Guest)
And oh, and we’ve got these things called success profiles. How do I use those? And so the functions that a hiring manager has to do currently within an environment, and what we were trying to do is go. There is a way we can take all of that pain away from you. We can give you a golden thread that goes from start to finish and then it’s over to you, but we can give you all the tools you need to do that job in a very easy way, and which is which is the tool that we’ve built, and all they have to do in, in simple terms, is give us a job title, tell us where they want to recruit and if there’s any special requirements for that role, and it literally goes into a screen on the ai tool that we’ve built and within the next two or three minutes, what? What it does?

11:10
It then starts to with the golden thread right through. It starts with that job advert, straight job description. We’ve plugged it. We put the large language model and what we’ve put in there is all the publicly facing data. So this is no private data and that’s really important.

11:27
We’ve put in our evp, our career site, old job descriptions. We put in all the civil service, commissioners, rules, all the governance from civil service. So it is very comprehensive what’s in there? And then, if we’ve got a couple of our directorates in in hmrc and we’ve got quite a few that have got agreed like ddat, so the digital data uh teams have got agreed specifications on their job, so we plug those in so it knows that if the job is in that sector, then it has to pull from that. And then what we really like and what’s been brilliant is that AI is pulling this information from the World Wide Web and everywhere else, but it’s pulling in the modern words, the modern job titles, the modern tools. So it knows if you put data analyst in, it knows the latest tools that people need to be using. And that’s one of the big changes we found that we’ve raised the quality of the applicants because they’ve got much better job descriptions and job adverts, because they’re asking for the current tools.

12:29 – Lucy Adams (Host)
How? How interesting, because that really matters to tech people, doesn’t it?

12:33 – Andy Headworth (Guest)
it does absolutely and and you’ll know from you know within hr, that how many, how many hiring managers over the years have kept that job description in the bottom left hand drawer and get it out once every five years and it’s totally out of date. Doesn’t work and so that that negates that. Yeah, and and so they get you know in a and it gives you the a full uh you know a full uh profile or advert.

12:58
It specifies is the. The specifies the criteria that you need to go against. It then produces a number of things. It recommends the success profiles that you should interview for and then why it provides interview questions. So it gives you a whole stack of interview questions again golden thread all the way through and you can build a secondary set of questions in as well. It provides you, then, with advertising, so it gives you the advert to place on LinkedIn if you wish, onto X or wherever you want to post it, and it produces all of this in one effector via Google Doc. You then come to the end of that, flip it into the Google Doc and then you can use AI or yourself to change, amend, improve anything that’s in this sort of effects of this PDF, and then you download it and you’ve got everything in one PDF. They can cut and paste the advert, they can do whatever they want. They’ve got the interview questions and it’s proving very, very successful.

13:56
We’ve got and we’ll talk about the other tool in a minute, but in total we’ve got way over 300 users in the business now using this tool as part of their recruitment and, anecdotally, the quality has been raised. They’ve had, in many cases, fewer candidates but better quality, which is great as far as we’re concerned. Because we get inundated, the time to hire is much quicker, because they don’t have to think about advertising, they don’t have to think about the, the interview questions, they can literally do that. Three minutes later they’ve got their document bang off, they go, they plug it into the, the recruitment platform, to post their job, and then they’re, they’re away, um, and just in terms of being current and being modern and having the reassurance that they are appealing to the right audience has made such a difference. So that’s the feedback we’re at. We’re still working on the metrics of the feedback, but overall, that’s what we’re seeing so far and it’s so positive to see that.

14:54 – Lucy Adams (Host)
I mean my experience experience. If you give leaders and managers something that’s going to make their life easier, they bite your hand off. Actually, adoption is is relatively quick. Has that been your experience?

15:06 – Andy Headworth (Guest)
most of them yeah, most of them they do, um, and there are some that we’ve had to encourage to use it, but, interestingly, it’s a little bit back to the days.

15:21 – Lucy Adams (Host)
Is that encouraging in terms of commerce?

15:23 – Andy Headworth (Guest)
Yeah, it’s a little bit back to the days of LinkedIn. If you remember when LinkedIn first arrived in, you know, early 2005 onwards, and you were doing sourcing with a hiring manager and you’d sit in front of a hiring manager and go, what are you looking for? And you put the job title in and all these search results come in and then you work through and they go oh my God, I didn’t know that you could do that. And what we’re finding with AI, it’s opening the eyes to people. Some of the younger ones and some of the more forward thinking are all over it.

15:50
But once people have used it and again I mentioned to you earlier before we came on the video unintended consequences we found out from doing this that a lot of them the biggest issue they had just interview questions. They didn’t know what to ask, especially with the different changing jobs and job titles and the way the market’s moving, and some of them just use it for the interview questions. Some of them use it for all of it and some of them just use it for the interview questions, some of them use it for all of it. So, again, you take a tool and they’re using it for the way that they want to use it, but that’s fine, that saves them time.

16:24 – Lucy Adams (Host)
Fantastic. Now you mentioned a second tool that you’ve been working on. Which pain point?

16:31 – Andy Headworth (Guest)
was that trying to tackle Insights, insights around where to hire, what to hire for so it was all about talent insights. So, from my perspective I own recruitment in HMRC and we’ve got, you know, 20 plus locations and if one of the customer groups are looking for, you know, to hire certain skill sets, we want to hire them in the right place across the UK, where those skills are or where they’re going to be prevalent from university, for example, in a few years time. And what I what I wanted was a tool that allowed me to know everything of an hour’s circumference, travel wise, from each of our locations. So I knew my talent location, everything from school school leave was right through to university, so we can get in there and do outreach through to, uh, current employment situation. So if people are making people redundant, or who’s hiring, who’s firing, and then all the skill insights that we we built into the large language model that tells us that everything’s going on, who’s hiring, what skills are there, what the salaries are, what the feeling is in that particular part, and so that’s what we built and we’ve got this tool and, again, very, very effective.

17:43
All you have to do you put a location in. It will give you an up-to-date snapshot what’s going on in that town or that city at that moment in time, again in a PDF, and the same way you would with ChatGPT, and allow you to question it and engage with it and talk to it. Then you can talk to this tool about that location. What’s the social mobility, what’s the D&I, what are the skills that are being hired, who’s hiring, who’s firing, what’s the salaries? And so it’s instantly.

18:12
It’s an instant tool for people to use and a very quick example of how much time we had recently had a commission in post change of government.

18:23
We are providing them with lots of information, as you might imagine, lots of requests and we had one to talk about different insights of the talent in all our locations with different questions and various things in all our locations with different questions and various things.

18:36
Now, before it would have taken us weeks, we would have had to go out and go to each location and Siobhan literally just had to put the prompt in, look at the locations, put the locations in and within 10, 15 minutes we had the table there with all the information on there for all 20-plus locations and we’re able to then share that on and that that’s a very, very simple, and we can go a lot more complex, but that’s a very simple example of the significant time saving and the ability to provide accurate data to make the right decisions of where to go is a really brilliant tool I want to have a chat now about introducing AI, new ways of doing things, forward thinking, progressive technology in a context sector, in your case the public sector, which is not I’m going to be as diplomatic as I can.

19:35 – Lucy Adams (Host)
It’s not known for its speed of adoption, its risk taking when it comes to and quite rightly so. I mean, you know, hmrc is dealing with highly sensitive personal information, etc. What’s it like trying to encourage perhaps a degree more curiosity, openness, speed of adoption in a sector like that?

20:17 – Andy Headworth (Guest)
And what, what works for you in terms of getting people comfortable, rather than multiple approval levels where you lose the will to live. Well, first of all, I’ve been very lucky. So in my role there, in the time I’ve been at HMRC, I’ve had, you know, I came in. You know, esther Wallington was a brilliant CPO From my perspective. She empowered me fully. She was very forward thinking and looking to the future.

20:33
And AI came along and it was like, yeah, you know, we talked about what we wanted to do and so I got the space first of all. So I got the space and the top cover too, and with Hannah up below her as well, to go and explore and see what we can do. So the first thing is I had that space. Other departments, sadly, don’t have that and they are regimented. The first thing is we deliberately picked tools to build that were based on public data, so the private data and people don’t have to worry about their tax data if we’re not going anywhere near that, yeah, and so the combination of those two things the license to the freedom to go and explore, and then, secondly, the the fact that it was public data really important and we use those.

21:14
And what we did really early on I was doing workshops and how we got people interested across the hr team and wider was to just talk about ai, talk about what it was about, how it was evolving, how quickly, how many presentations I rewrote that year as it changed, I do not know and we’re just doing just gentle workshops with people and getting them to see what it could do, have a bit of fun, but also explain what a prompt was, explain how it would work, explain how it would work in HR. And then we did that across HR and then other customer groups and business units said, oh, can we have that as well? And then they’d go off and play. And then we had wider area across the civil service had heard about it and so can you come and do it over here? And can you come and do it here? And so I ended up going to do this. So we were basically starting to get people stage one, get the interest, get them playing with it.

22:05
If you had, we were lucky in hmrc. We’ve always had access on our system to all the tools like chat, gpt, uh, yeah, and of course some, some organizations don’t, do they?

22:15 – Lucy Adams (Host)
they, it’s closed off to them.

22:17 – Andy Headworth (Guest)
But they do have a mobile phone and all those have been on the mobile phone and they can go and do it on there. So we were always saying so when I was with other departments that were maybe locked down they were, you know, and the answer was simple Go to your phone, just type in chat, gpt, log in with Google or whatever it was, and just play, just ask it something, just think about it, ask it about HR policies, hr strategies, talent strategies, and start to see how it works and start to see how it. And that’s what we did, and we just did that for the first year. And then we put in a bid.

22:53
Civil Service Live is a global event across the UK for all the wider civil service and myself and a couple of colleagues uh well, they led it, uh, mark and siobhan and we were invited to go and present to all these big hubs and there were there were five, six, seven, many more thousands of people at each one from all the departments, um, and we did this for about three months, going around the country, for literally from edinburgh to newport to london, um, and what was interesting was, across all the talks that were done and there were hundreds of them. We came second in terms of the most interesting talk for people that they wanted to explore, and so this is just everybody in the civil service. And we talked about curiosity. We talked about how, what AI is for you as a person not necessarily in a workspace and what it means to you and how you could engage and how you should think about doing it.

23:47
So it was very gentle, gentle and curiosity was the one word that we really focused on Be curious, go and explore, don’t be afraid. You can’t make a mistake with this. You can’t as long as you don’t upload personal data and don’t upload company data. And you just keep that in mind. And the thing that we said from the start and I think it still applies many organizations have got a social media policy and, if we think about it, this is very similar to a social media policy don’t be silly, don’t put data in there, don’t put company data in there. And if you think of and that’s the guidance we gave in the early stages to say, adopt that mindset and you won’t go far wrong, yeah, what?

24:28 – Lucy Adams (Host)
what do you think are the the common misconceptions about ai that you come across? I’ll give you. I’ll give you an example. So when I first started using chat gpt, I used it like google. I used it as a, as a source of information, um, as opposed to something that can help me create things, and that was a kind of revelation to me. Is that a typical misconception that people have? Are there others that perhaps frighten them about using it?

25:01 – Andy Headworth (Guest)
Well, we haven’t. We haven’t touched on ethics and bias, and that that’s going to be the two things that people always go on about. Bias, the first thing they will say is biased. It’s always because they read it in the Daily Mail and everything else that we sorry Daily Mail, but that’s true. It talks about bias and other things. So bias, unfair, I think, the two and the ethics of should we be using AI? So we’re in the HR space. We ultimately are going to own ethics. Really, I think people, ethics and how it’s used that’s to be explored, I guess. But it’s a good question because everyone’s different and I had some fun. If people want to look at this, it’s in my LinkedIn profile. Right scroll down. We’ll put your LinkedIn profile link in the show notes.

25:54
So how I did this in a in I did this example to show people how to do it differently in presentations. So what I did was I, what’s brilliant about things like this and you can create your own virtual teams and world and whatever. So I set up a, a group of people. So I I asked chachi pt to have, so we’ll have a uh, a chair person of a panel, and in that panel I want this person, this person, this person, this person gave a small persona of each of those people in the panel, so I picked a random subject. So just to prove a point should you put cream on a scone first, or should you put jam on a scone first? So as part of my panel, Everyone knows it’s jam first.

26:35
Yeah, well, not in Devon. And so basically I created this and then asked ChatGPT to explore this, debate it and then provide the output, and which it did and it copped out and said you can have either. But it showed and how it works and a simple analogy to something that’s not related to work and people. Oh, that’s clever, that’s really clever. Oh, I can use it for this. Then I can, so I could set up a little marketing panel to test our marketing strategy, and those personas could be the ones that I think you’re absolutely right.

27:11 – Lucy Adams (Host)
I think too often what we hear about AI is big stuff you know it’s going to take away this many jobs, or it’s going to impact this many industries. Or we see fake you know photographs and videos and the scary stuff but actually just on a very small and personal and human level. What could you do with it? That’s fun, that’s interesting, that you can help your kids, you know there’s, I think so, to reduce it down a little bit away from big AI into actually something that this is, something that is for everyone, and the more we can do that, I think that’s that’s useful. I’m just going to finish off with a couple of questions. One which is around you raised it earlier about inclusivity, bias and what’s your response to people who have anxieties about inherent or unconscious bias being hardwired forgive the language, hardwired, you know what I mean into AI. What’s your response to that? What are the solutions to that?

28:20 – Andy Headworth (Guest)
I think, first of all, the main three providers behind Perplexity, claude and OpenAI, which is ChatGPT, are working very hard to eradicate that. So that’s, if we think back two years ago, it’s a lot better than it was then. So in the big scheme of things, but when we’re building your own within an organization and you’re building your own large language model, you’re in control of that data. Now, what we found, that’s been interesting. So, straight as an example answer to that, you know we are looking at building a tool to help so that. So the biggest challenge, the number one challenge we’re trying to solve, is how do we help hiring managers or the business go through the large numbers of applications that come in and reduce that to a manageable number? We’re talking tens of thousands we get, so we have to, and we usually do that manually, so you can imagine the time saving. So we’ve done some simple testing with completely anonymized data at the moment and we found that you can build a prompt. Obviously, it’s a complex prompt and you can build a prompt in a certain way and construct it and you can challenge it, get it to challenge itself around bias, around fairness and around inclusivity and, for example, we put reasonable adjustments, which is something in the civil service that allows people to be able to go to interviews and have conversations, allows people to be able to go to interviews and have conversations. So we built all this into the prompt and then ran a whole load of CVs through it again, completely anonymized. So there’s no data there, and what we found was because we had the data from those original interviews, so we know the outputs.

30:01
What happened in real life with non-AI? We found out that this is not going to surprise anybody, but humans were more biased than the AI. It was because we’d built this tool and it questions itself and looks at the different words that are used, and so you can build it to be as unbiased. But the bottom line is we live in a biased world and every single one of us have unconscious bias in us. We try and eradicate it at interview, don’t we? But everybody’s got some, and I think that we found when doing this last exercise.

30:34
It’s consistent. So the thing about ai is it looks at it independently, it looks at the data, it doesn’t look at anything else, and so it’s examining the data against another set of data, but it’s doing the same thing every time, consistently, and it doesn’t. So, whatever you know, if there is a lower level of bias in there somewhere which is inherent, it’s very low level, but it looks at every single person or every single bit of data the same way, every single time, whether it’s 50 ways or a thousand ways. And so when we’ve shown people this and people see this, and we’ve examined it afterwards and compared data sets, we can we can prove that it’s better, but it’s always going to be there. People think again. People hear the word bias, um, because it’s a favorite word across and, as you say, you’ve got two things in its favor.

31:24 – Lucy Adams (Host)
One it’s you know it can. You can challenge it over, and, over and over again to remove the biases that are within there, but once you’ve got it as good as you probably think you can get it, it will behave like that consistently, unlike a human being. Final question for me, then, andy what’s next for you? What’s your next area to explore?

31:45 – Andy Headworth (Guest)
What’s your next area to explore? Well, we’re referring to at the moment, which is that we’re trying to get the magic around sifting and selection and assessment.

31:53
So that’s the next one, that’s the big one for us because it’s such a time saver and it would make such a difference to the organisation and the wider organisation.

32:02
We’re currently going through all the governance boards, the ethics boards boards, the ethics boards, the data boards. So that’s our current process because there is people, data involved. But just to be clear for anyone listening, we are not looking for ai to make decisions. What we want to do is, when we get a substantial amount of applications, a large proportion of those are unsuitable and what you know you touched on earlier can use for so many different ways. It’s very it can very quickly look at, look at things and go, no, it’s not suitable because it doesn’t meet the specification, and then obviously we use AI to manage that. So we’re looking to do it in the right ethical, transparent way and we’re just going through all that process at the moment. So we have a few few. We have you mentioned governance earlier. We have some governance. Now we’re working through um, which is what we do. So hopefully the new year we’ll see that.

32:55 – Lucy Adams (Host)
We’ll see, uh, we’ll see the next iteration, andy. Thank you so much for joining me today on hr disrupted. As I say, we’ll put the link to your linkedin profile in the notes on the show so people can connect with you if they want to find out more. But for now, thank you so much.

33:09 – Andy Headworth (Guest)
Thanks, Lucy Pleasure.