What New AI Skills Every HR Leaders Needs in 2025
In this episode of the HR Leaders Podcast, we welcome Stuart Martin, Director of Global People Operations and Services at Lloyds Banking Group, to discuss the role of AI in reimagining work and transforming HR operations.
Stuart dives into how generative AI is reshaping employee experiences, automating tasks, and addressing skill shortages. He also highlights the importance of creating scalable, agile systems to enhance productivity and personalization in the workplace.
🎓 In this episode, Stuart discusses:
Why HR needs to rethink outdated operating models to leverage AI's potential
The role of new skills like prompt engineering and automation in HR transformation
How generative AI is reshaping workflows, task automation, and employee services
The importance of balancing automation with human empathy for moments that matter
How Lloyds Banking Group is deploying generative AI across 65,000 employees by 2025
Deel is the all-in-one payroll and HR platform for global teams.
Deel helps companies simplify every aspect of managing a workforce, from onboarding, compliance and performance management, to global payroll, HRIS and immigration support.
Deel works for full-time employees and independent contractors in more than 150 countries, compliantly.
And getting set up takes just a few minutes.
Chris Rainey 0:00
Stuart, welcome to the show. How are you? I'm very well, good. Thanks for inviting me. Thank you for getting through all the snow and the rain. Our first for the season. I think that's the first time with snow this year, right? It has it. Yeah, settle. Though. I was thinking, Oh, my daughter's probably thinking, please, please let the snow settle.
Stuart Martin 0:17
Well, my daughter's in Oxford, so she probably has settled. She sent a pic this morning as she got up. Oh, really settled. Sort of very sort of picture postcard kind
Chris Rainey 0:25
of, yeah, we don't really get that in London. It just sort of melts, then goes into slush, I thought, the most glamorous way how you been
Stuart Martin 0:33
very well. Thank you. It's been a busy period, some international travel so and some of the thinking I picked up in Paris, where I was at the unleash conference, I will, I'll certainly come through in our discussion today, ai, ai, Gen AI, or every session, every session. But that was really, really good, some really good sessions, good speakers. And then we have an international business in India, so in Hyderabad for two weeks, just working with the team there on some nice, some heavy lifting to be done. And then it's the season for annual you know, look back, look forward, 25 strategy stuff. So that's it's really keeping me occupied. And then we have a full on run till December 20.
Chris Rainey 1:15
Amazing. It crazy. Already feels like the year is done. Yeah, as soon as you get to, like, this period of the year, it's like, it's almost like people start winding down already, right? But it's also crazy busy at the same time, yeah, as well, tell everyone a bit about your role, because you have a quite unique role responsibility as well, and break that down for everyone. Yeah,
Stuart Martin 1:36
the title is a bit misleading, but also a bit of a hint at what I do so at the so I'm part of the HR Exec. I report into the CHRO Sharon Doherty. We are in a Lloyd's Banking Group. We are a large scale, multi year digital transformation, one of the largest in financial services. If we then break that down into the HR space, this way, my role is probably key. I'm responsible at the macro level. I'm responsible for ensuring our colleague experience is fully integrated digital we enable all the emerging technologies so AI, and we'll talk about that a bit later. I'm sure Gen AI is what we're going to introduce in December. But ultimately, it's about getting a comprehensive end to end, very modern digital colleague experience across the bank for all our colleagues. And the three component parts, one would be HR operations. So, you know, big focus on making it almost zero touch where we can with huge amounts of automation. The other bit is I also look after what's called a conduct team, so any issues, concerns that come out of the out of our branch and network, etc. We have over 65,000 colleagues. We help our colleagues through that, from wellness to absenteeism. And then I manage what's called the people platform is which is a fairly recent capability, two years in two and a half years in so we have across the bank multiple platforms. I look after the people platform, which is essentially a combination of our tech change process, colleague journey capability in an agile capability. So I'm really quite fortunate in that space, and get gives us a great agility in the market as we face off to our colleagues, we have a backlog into 2026, for example. But that's, that's, that's the capability we in place. So very agile. Does that include
Chris Rainey 3:30
people analytics as well? Sorry, does that include people analytics? When you mentioned the technology,
Stuart Martin 3:34
we've got a little bit of it. It's still the people. Truthfully, our people analytics capability is emerging. So in the HR op space, is where I'm building it. We also have what we call the data hub. So there's another one of my colleagues manages that for various reasons, but yeah, we pull on that as well. Yeah. So it's an end to end capability, and we're growing that out. Was it was
Chris Rainey 3:56
the team intentional about making sure that you look after those three areas, because I think it's really important, because many companies, they're quite discon, they're quite disconnected those three areas you outline the fact that you oversee all of them. Is that intentional by you and the team? It's but
Stuart Martin 4:08
it's by design, yes. So the platform has been an incredibly powerful capability. As I said, it allows us to build a backlog of tech and transformative change we no longer have change managers, for example. We no longer have radio managers. Know, in an agile way, we have some at the one level, down in the platform. We have a business platform lead, very senior individual. We have a technology platform lead, very senior individual. They work collaboratively on the tech and change. We have labs. We have lab leaders. We have products a lot of owners. We have customer journey managers. So we've embedded design thinking in that capabilities. So you know, fundamentally different culture change as well. Yep, yep. So in some cases, we've managed to. Up skill and cross skill, sort of traditional tech and change people. But truthfully, we've also had to bring in quite a few external individuals who have a very different background in mindset. So that's come through quite strongly. So in the in the platform space, about 350 colleagues in that space, half of them have been in the role for 12 months or less. Wow. Some of that external, some of them bringing it from other parts of the business, and some of them we've up skilled through a very rigorous assessment process. So we're doing a huge talent refresh as we head into this
Chris Rainey 5:34
space. Interesting, yeah, but it kind of leads us to one of our questions, because one of the things that we're hearing, you know, is AI going to take our jobs both in HR and outside HR. You think that's a fact, or do you think it's a myth?
Stuart Martin 5:47
It's a great question. Thanks. I think it's probably somewhere in between, in the way, the way I think about and from what I've seen in the space in the last few last while, is that it's already a fundamental part of how we operate. So I make a distinction about what I call embedded AI. So most of us have smartphones as they already that's embedded in the way we work. I know people wear watchers. They wear rings that, you know, various
Chris Rainey 6:14
I saw your watch exactly. Yeah, and the
Stuart Martin 6:18
ring, I mean, the ring, there's the arrow, the aura ring, with, you know, one of the products. And you know, that's all using AI as a way. So it's part of our personal lives in many, many ways, and huge advantages and benefits there within the workspace, exactly, you know, very similar. I can pick many examples, but let's talk about HR. So we are using embedded HR for a good few years in Louis, we went on to work day as our cloud solution in 2018 and we've continued to expand that. We have service now, and I'll talk about talk more about that in a few minutes. We have Microsoft and all the Viva products. So embedded AI has been part of what we do for a long, long time, and, well, good few years. And
Chris Rainey 7:00
is that just be clear for the audience, though, is that that's not generative AI, not yet. That's AI. Okay, let me clear it, because you mean that may be obvious to us, but I think to all it's maybe a bit confusing. AI has been around for a while, a long while, but then we're talking about generative
Stuart Martin 7:16
AI. It's a good clarification, and I'll come to that in a sec. But I know the point is AI has been then, and it's doing a lot of work, yeah, on the Gen AI space, which is more so two, two and a half years, commercially available, although technology has been around for a while, that I think is going to have a much more transformative impact on on work. And I think in that space, it will certainly take over multiple activities of work that's been done by humans in today, and I don't think it's going to take over people's roles. So we are using it quite extensively. We do case summarization, we generate various documents, etc, using the Gen AI capabilities, so that that's playing into that space more and more. So it will fundamentally either change or completely take over activities and tasks that people are performing. The upside or the flip side to that, though, is it will force us as a workforce, as individuals, as humans, to play in a different space. So Gen AI is not creative. Some may debate that, but you know, it doesn't have a creative capability. It struggles with empathy. So, you know, it will, I think it will unleash and force us to be much more focused on creativity, collaboration, human interaction, some more innovative ideas, etc, etc, which we continue to leverage Gen AI to help us drive that. But I think that's the space we're going to play in. So I don't, you know, fundamentally, I don't think it will take over jobs. And, you know, we would all be in a sort of catastrophic situation for that reason. And then just, I kind of think a little bit broader about it, you know, the social demographics are there. So we have an aging population. We simply have two people, too few individuals coming through. So that's that's gonna help, that's going to be a help, help us quite a bit. And then we have a Global Skills deficit, you know, we have a Global Skills effort, and get into more that, if you'd like. But essentially, we're not going to hire ourselves out of this, this, this challenge we're facing today at the workforce in an HR level. So I don't think, I think we'll change the way we operate. Will change the work we do, but I don't think it's going to become bit of a distinct environment.
Chris Rainey 9:25
I love your point, by the way, about the creative side, it's only I've been thinking about a lot. I think now you can code using AI, you can, you know, finance is pretty much taken care of. In most cases, you can do a lot of the finance work. So it's really like what you kind of a lot of the analytical skills, and then when you look back, it's like, okay, now really it's the creative piece. Becomes adjustment that more important as well. Yes, you can create stuff with the platform, but again, you have to be have a creative mind. Yeah, to be able to do that, one of my worries and concerns. Is because you have all of these tools. Is that going to eat away at creativity? Because I can, I can, I can just write a prompt and create a video or a picture or a piece of art, you know, that's a different set of skills, of what it takes to paint portrait or draw a picture. Does that make sense as well? So like, when maybe the new generation, when I used to have to write a script for the show or write a video script for something I'm thinking about because I had a blank slate, I didn't have generative AI. So it's in the process of me writing that I learned a lot about myself. I developed those skills, right? If that makes sense to storytelling. But now with chatgpt, you don't have to do that. You just say, hey, write me a story around XYZ. Boom. You've not had to go through the creative process in your mind, if that makes sense, to bring that to life. So I'm worried about how that's going to impact,
Stuart Martin 10:57
yeah, and I'm going to add insights to that, if you don't mind. So I think from a creative perspective, it's going to be a challenge. My thoughts are slightly different. I think we'd have to engage different parts of our brains and to be very and be creative in a different way. I think there'll always be people like yourselves and others who are, you know, artists do watercolors sketch, and they will create a something unique and specific, and there'll be others who are probably in the space already or and as younger generations for them, this would just be part of the how they operate, yeah? So where, you know, they're gonna have to think of what's the creative prompt? Yeah, great about what's wrong? Yeah. Creative prompt. You know, we get into the space about writing an essay or writing a piece of tech, doing a piece of technology. I mean, do we really think that? So, I think we're going to be much more we need, need to be much more creative on the code that we write to get a different result. And then my thoughts are on insights. I think we will have to really amplify our ability as humans to really see trends and insights and say, well, actually this, this is something unique I hadn't seen before. Yeah, the other space around you know, creativity. We go through the medical field, for example, we are creating, we are looking at and generating cures for so many illnesses at a rapid pace, great for me. There's the there's the bar, that's the cure, then the inside of how do we apply it? How do we take it to a global audience, etc. I think the other spaces that are probably more creative that we haven't had a chance as to play in before. In the artistic space, there are various things you can do as well. So, so I it's a good point. I mean, slight tangent, if you don't mind, I do have a slight concern, which I'm sure we'll find an answer for as well, and that is, as over time as Gen AI becomes more prevalent in our lives, completes the task and activities we've just spoken about. You know, will we get get to a place where some work is pretty dull and boring?
Chris Rainey 13:06
Yeah, that's part of it, yeah. And
Stuart Martin 13:09
then what do we do? Because we're not gonna get free time. I don't know the answer to that one yet, if I'm absolutely honest, but I there's a part of me which believes in human nature and creativity, and I think we'll find the answer in that space. Yeah,
Chris Rainey 13:21
because I think it's for me, it's like, it's in, it's within that exploration that keeps me engaged and excited. If that makes sense, when it's sort of given to me after a quick prompt, it feels very transactional and boring to your point, to your to your point. So it's gonna be interesting. We're so early in it's now, like for me, it's a day to day activity. Even just before we sat down here, I ran some of the questions through a agent that I built and to give me some feedback on anything else that we should be asking. So like for me, it's just part of my flow of work already, but I still take time to not use chat GBT as well, and actually just sort of idea, right, literally on a piece of paper and write stuff down, because I want to keep that creative juice going, and don't want to just put all of it into my agent, if that makes sense. So I'm trying to find a balance.
Stuart Martin 14:13
Yeah. I mean, I heard someone I was fortunate enough to be in a session last week with Lord Holmes, who's really a huge proponent of getting a regulatory framework in place for for AI and Gen AI in the UK. And he made, you know, he made a comment which I've kind of think about right now, which is that it's a bit like electricity came along. It's a bit like the railways came along. You know, only 40 years ago, you know, World Wide Web came along somewhere in there as well. And those were all really, really massive, transformative changes to society. And yes, we left some things behind, but we've, you know, we've, we've really gained so much more. And I think we in that early phase there. So there's a there's a human there's an optimistic part of me. We still, we had the same thing, right?
Chris Rainey 14:57
When the internet came along, everyone was like, everyone's. In their jobs, right? It was the same. It was the same. It was the same thing. Same. Hype on the skills piece. How do you think this is going to change HR operating models, and what are some of the skills that HR leaders need in this sort of post AI world?
Stuart Martin 15:20
They're probably multiple parts of that question, if you don't mind, I'll try and break it down, but if I go off on a tangent, pull me back. So my first thought is, as we've just been talking, I'm really excited to be at a sort of phase in our in sort of way of work where we can actually reimagine how we do work, but which, I mean, you know, if you go back 10 or 15 years ago, we did process engineering, and then before that was Lean Six and we are, there were lots of ways where we really, in my mind, we were just tweaking it and doing things different way. I think the value and the beauty of generative AI is that it allows us to kind of step back and reimagine how we do work, where we do work, etc, etc. So for me, there's a interesting point in the way in sort of world of work, where this is going to be transformative. So I'm excited about the opportunity to reimagine work. So that's kind of my first point. And then to answer your question a bit more specifically. You know, what does it mean from an HR op model and HR perspective? So an OP model perspective, I think all the thinking and where we set up of the last, you know, 2030, 4060, years for the industrial age, and even for the knowledge age, we're going to have to question, and that's going to move away. So let's break it down to a much more specific example. We are at Lloyds going we're introducing next month agenda of AI across all our colleague experiences. So as you interact with us, we'll have a Gen AI capability into and in support of that. We've moved away from the traditional way of setting up a a contact center capability. We have far fewer people in that space. We are targeting over 80% of queries coming in to be answered by the virtual agent. We are building some empathetic capability so that the response is there. We are looking at when a colleague enters a prompt, if they have a search capability, if there are certain keywords where we believe we need to intervene, we will triage that immediately to a to a human agent. And so there's some you know, emotional someone has well being, for example, well being, or I'm not feeling too good today, or how do I deal with depression? So we've identified some keywords, and we'll continue building that. We triage that immediately, but the rest of it, we think we've got a high deflection rate in place, and that's how we're moving. And so the skills I now need in that contact center capability is very different. We don't take calls. You know, it's live chat. Calls are very small. I need individuals who can look at all the queries coming through in a day, build their own workflows in the flow, not go to it, wait for pipe term. You can look literally in the moment. You can look at what the workflow issues are. I can build my own workflow. It allows me to say, well, actually, I see there's a trend. There's a frequent asked question about what's the impact of the legislation changes, which was announced in the budget this morning. I'll make this up. You know, we can have some frequent ask questions. That's fairly bog standard, but if I notice a trend in the queries coming through, I can build the workflow is actually Chris, you know, this is the impact of it. Here's a scenario from a tax perspective, your scenario from your benefits perspective, etc, etc. And it becomes highly personalized journey for you to really understand within a day or two what the implications are for you in that workspace. That's a very different skill set, operating model and
Chris Rainey 19:01
capability, yeah, that still didn't exist in the function. So then
Stuart Martin 19:05
I'll go down to a few more specific skills that I see we need,
Chris Rainey 19:09
which is, sort of, how do you define that as a skill, though?
Stuart Martin 19:13
So today, we don't have people sitting in a contact center tracking all those queries coming through, and then have the ability, because we've got low code, no code, and we've got automated workflows to build those workflows. So that's a very different, yeah, that would have been, you know, written down. We've had two workshops, you know, and three months later, pass it on. Thank God. They go, it happens in the moment, and answer your question in a slightly different way. So if you had said to me 18 months ago, within the HR operational space, I will probably need people who are pretty skilled at prompt and prompt engineering, didn't know the term, what are we talking about? Surely that's an IT skill. What. Yesterday, as you know, as I, as I set up, as I mentioned earlier, we will have people who will need to understand prompt, prompt engineering capability. They will need to be able to build some automated workflows. And let's be clear, it's, it's almost click and drag it now it
Chris Rainey 20:16
is, yeah, yeah. But building workflows do have to be still, have to have the you start to understand it in order to create a workflow. Yes, right.
Stuart Martin 20:24
And that's a very different profile than AI find. Ai protected up sort of 2345, months ago. One of the other things that we're doing, and I've gotten a bit broad in this, you know, we are bringing product ownership of colleague, facing products into the into into the operation center, one example, and there are many others, but one we are using is the Viva product suite. We deployed Viva learning into our colleague environment earlier this year, and the updates, upgrades, changes, small tweaks, functions and features to make it even more personalized. We will own within the HR space, because the tool itself help is that intuitive, and we wouldn't want to have to go back into it in the IT shop, go into their backlog and get all done. We will build that capability internally within HR operations to make small, configurable changes with huge guard rails in place are sort of on the go. So we'll have small within a week, and have 20 or 30 small tweaks to the learning platform, which wouldn't have done a year or two ago. And that's a very different skill set. I need to understand the product. I need to understand the implications of making a change, etc, etc. So those are very different skills and capabilities that we're seeing in operations as we go and then the final example, we also moving on to a cloud based a cloud based Payroll Solution. We've lagged a little bit in that space. Early next year, we'll be on a fully end to end. Cloud solution and the capabilities I need to manage the run of the payroll operation is very different. I don't need people now to check to do checks of spreadsheets. We're moving from seven payroll solutions into one. Wow. And so the capability and how long was that taken? We started October last year. Okay, pretty good, pretty good. So we'll be live grass in February next year. Yes, pretty good, good. But again, I don't need a large number of people. I don't need people. Excuse me. You do it with manual checks, balances, etc. They need to read the technology and move to, you know, more insight. What did
Chris Rainey 22:41
you say earlier, the percentage of all of this then of people that you're upskilling versus new
Stuart Martin 22:46
so in my part of the world, we have managed to upskill and cross skill 49% of the team in the past 12
Chris Rainey 22:56
months. Nice. That makes sense. One of the things I'm hearing from this, which I think is really important for the audience. Takeaway is you can't just add generative AI to the current HR operating model, and that's what I'm seeing companies do, and expect that to work. You completely re imagined that, and everything you said, I kind of took away. You've got speed, you've got agility with all of those. And in order to do that, you've had to start from scratch, almost, right, and kind of completely reimagine again the old industrial revolution model that came in is not gonna you know what commercial goals say, What Got You Here Won't Get You There. But I'm seeing so many companies make that mistake where they're just sort of adding this technology to an old, outdated HR operating model, and they're like, why are we stuck? Yeah, right. Why can't we move? And you've done two things. You've kind of re imagined that and built that out, obviously, with different partners. And you've also had to go through the upskilling and re skilling, and there's no easy way of doing either, either one of those. I want to touch on the rollout of generative AI, because there's not many companies that are in the position that you are HR space, what vendor you doing that with? We're using service now. Service now, yeah, and they too, and that's what it's for the it's the the the agent that you're rolling out, yeah, yeah. I can talk about that a bit more. Yeah. I love to, because there's not many companies that are have that level of AI maturity that are rolling it out. Many are in the conversation as well, and this isn't a plug for ServiceNow, but why service now and talk me through the process you had to go through internally? Because even with Atlas copilot, our own HR copilot, we're building, we're seeing that internally, it's a challenge to navigate in terms of getting the right stakeholders involved, the right buy in, you know, the IT security implications. Talk me through how you kind of navigated that, and then we'll talk a bit more about roll out.
Stuart Martin 24:50
Yes. So let's
Chris Rainey 24:54
talk about product choice, yeah. Firstly, because there's a challenge for our audience. Honestly, they're inundated every single, you know, every one of their. Vendors I was speaking to, I can't say the name, yeah, I see it forever, a very big company, because he would tell me off if I said his name. And he's like, Chris, if I wanted to tomorrow, I could turn on seven agents, yeah? Because every single one of my vendors has an agent now, right? And it's now, it's have to be very conscious about how we make that choice, and that's a big challenge, so I'd love you to speak to
Stuart Martin 25:17
that. Oh, absolutely not, in our face the same challenge. Even though we've made the product decision, we still get, you know, everyone wanting to tell us that theirs is slightly better. So I take the challenge, and it's a strategic choice, really. So let me tell you our scenario, and hopefully others can learn from this, or, you know, at help with their thinking. So first and foremost, our system of record is work day, and that's, that's, that's, and we in a very fortunate position to have that as our system of record. And that's end to end. We use all the extensions, etc. The choice are, the rationale on the choice for service now is to the Gen AI capability was driven partly, mostly because we've got some of the virtual agents in use in various parts of the HR function. Okay, it's sort of, but it's all sort of here and there, quite fragmented, so we have a sense of it, and we started using it, sort of piloting some of this capability, earlier this year. The second big reason was, as part of our enterprise agreement. We use service now as a tool across the enterprise. So that gives us commercial leverage. It gives us a different relationship with ServiceNow from that perspective. So that's been a big plus for us. Some of the other rationale has been if you play with well known vendors who have a global reach, or at least a regional reach, especially with this new technology, it gives you some level of confidence. So some of the built in elements are on guard rails, access to data, where the data gets used, how it gets archived, etc. Data privacy challenges, when you play, when you have a big vendor, a renowned vendor, that gives you that level of confidence. If I turn specifically to the decision around the use of the Gen AI capabilities. So this is the now assist. What's the name? By the way, I know you can now now. Assist is the tool we're going to do a big reveal and an internal we ran an internal competition recently. We got people kind of nominate names, and we at the point of making the so on. Is this
Chris Rainey 27:35
gonna go out? You said? Did you say earlier? 80% of the employees? You'd say
Stuart Martin 27:39
so by June next year, it will be available for all our colleagues. Okay? And we have 65,000 colleagues in multiple locations, including India. So that's, that's the plan, in December. So next month, we will move out of pilot and start scaling. So we'll have a seven month scale. What does pilot consider? How many people is that? In fact, about 500 people. I want to be playing a tool, building up some of the elements around large language model capability, checking for naughty questions. So you know, so we can people deliberately to ask questions that we are have a response to think about training agent. So, yeah, start training agent. But also just getting our mindset right. So there's a data build, but to your to your earlier question. So, you know, if you have a vendor, you know, make a choice, but having a global player gives you that reaching capability. And there's a, there's a another element which I thought, which I think will be useful, uh, certainly worth sharing. One of the things we signed up to was that we knew the Gen AI tool now Assist, which is the service now, one is a new product, new in the market, new in the space, being used in some other parts or functions, but certainly new in HR. So we've deliberately agreed and signed up to what I call a co design concept. So we've said to now assist, let's work collaboratively on introducing this into our workforce. Don't go into a dark cupboard and, you know, come back and say, ta, da, you know, we will off the shelf. Yeah, yeah, no, no, off the shelf stuff. Let's do this jointly. Let's work collaboratively. Let's do a bit of CO design. It builds their muscles in a different way of operating. And so it does with us. And so from our perspective, you know, we have little to no interest in the IP what they do with it is theirs. But we want a very we want to get to a place where the product works for us, and rather that, and so that CO design, co collaboration capabilities is proved to be incredibly helpful for us. So if I were to call out something in particular, go for the bigger player, a known player. Don't go niche, because then you really struggle, and you can get in sort of a co design space that would be very, very harmful. Those would be the three, three or four big ticket items, from my
Chris Rainey 29:44
perspective, who needed to be involved in that decision making process to get this over the line? Because that's again, challenge with our audience. Yeah.
Stuart Martin 29:50
So we know rather fortunate position in Lloyds, and that we want to get in the space very early because of our large scale digital transformation, we did an exercise and. Um, this time last year, into early this year, where we went around all the functions in the bank and said, This is what we want to do. What are the opportunities to leverage Gen AI in a safe, reliable, transparent, non biased way? And then got a a backlog built up more that we could tackle. And then we went through a prioritization exercise so this would be cross functional with our IT organization, with our data organization. We also had, we also have, in one of the first in the FS space, financial services space, to have a data and Artificial Intelligence Committee, which is cross functional. No
Chris Rainey 30:37
surprise, you can't you literally cannot roll something like this out without having that presented,
Stuart Martin 30:41
I represented, obviously, a child, or people, as we call it, in Lloyds with risk finance, we had you
Chris Rainey 30:49
bring that to the table. Did you? Did you said that I was always interested to know who
Stuart Martin 30:53
we have, a chief data officer. And that was, that was amazing. That was east. Okay,
Chris Rainey 30:56
great. Everyone has a chief data officer. Great with the capability
Stuart Martin 31:00
where risk, you know? Yeah, I think we had 14 different capabilities represented in terms of that incredible and that that meets every month,
Chris Rainey 31:10
once a month. Would you other companies call it AI, Ethics Council, Yeah, same
Stuart Martin 31:16
thing, data, artificial intelligence and ethics committees love it
Chris Rainey 31:23
a bit more easier what's
Stuart Martin 31:26
happening out to regulation perspective, we look what's happening love that EU is a way to kind of benchmark ourselves what other competitors are doing, and then we look internally and what what we're doing. So in the HR space within Lloyds, we are the second function to use Gen AI. So we're using it internally in our
Chris Rainey 31:43
with our first finance
Stuart Martin 31:46
customer relations team to help them answer clients Yeah, then we don't have any client facing within your Lloyd's app within Yeah. Oh, that's amazing. So, so so we have providing them. So we're providing Gen AI capability to be able to read a customer complaint, look what the thread is, provide them a suggestion, commentary, but there's still the insight. So they don't just do a read, don't just do a showing. So what they have to read? And so, so that's, that's the first place. And then we, can we be second? As I said earlier, the last few months, we've been playing with family and friends, so in small proof of concept. And then from December, we will start scaling what's
Chris Rainey 32:23
some people may not be aware of, like, what service now, what content, what assets live in there that can help employees. What? What is it specifically with that connection with ServiceNow that What? What? So what can surface to help people? Because, yeah, not everyone knows what that is, yeah,
Stuart Martin 32:39
yeah. So we're going to scale, as I mentioned earlier. So there are two elements to to that answer. So we'll scale to all 65,000 employees by June next year. That's the plan. So volume scale, so it gives you that volume and complex. Our objective is quite clear. We want to we want to enable our colleagues to whether they're on their phone or any other company device, be able to access HR information? Well, break that down, though, yeah, I'm going to do that. Okay, great. I was gonna go there. And the first thing we're going to do is creative management. So we have over three and a half 1000 policies in the bank. We are 300 plus organized, 300 year old organization, and we built up quite a lot of collateral. So it's about being able to access highly relevant policy information in the first instance. And then we'll get that more complex. I'll talk about that in a second. But we know from some research that we've done internally, our colleagues can spend five to 10 minutes looking at any for doing search for any policy. If they're really lucky, they'll find it. If they're more lucky, they'll find the most relevant one. So what the analysis capability is going to enable us to do is provide it to you based on your role, based on your location, based on your grade. So highly personalized, and you'll find the most relevant policy content. We also changing the way we present that content from very deep, lengthy, text based, capable, you know where we stored it to much more with icons. It takes you through hyperlinks very quickly to what you want to do, etc, etc. So it's a much more relevant, personalized experience. Our medium term aim is to make that conversational, but let's get there. When we get there right now, it's search based. And then to answer your question a bit a bit further, as we go into next year, we also need start doing transactions. So I can look at, I'll play through an example. I can look at, what's the paternity policy? Yeah, it'll tell me exactly what it is being on my circumstances. We will then get to the transaction bit, which is to say, actually, Chris, do you want to apply for I see you looking it up. You can. Do you want to apply for it? Yes. Do I qualify? Yep. What's the dates you wouldn't do you entered in and off you go. So then updates work day, does the workflow to your manager. And you know, that's all done. Within, within, within a minute, it then will say, Chris, do you want me to populate your Outlook calendar with the dates that you're going to be off? Yes, boom, you know, and that's all sort of seamless. Do you want to set up a reminder for you and your manager to to, you know, a week beforehand, or two weeks beforehand to say you need to start the transition. Yep, we'll pop a reminder in your calendar. By the way, Chris, you know there is a checklist for people going on paternity leave, some of which is internal to Lloyds and someone's from the government around some of the benefits you can apply for, etc, etc. And here's the checklist, and I'll email it to you at the end of this conversation. I love that. So all of that just happens, you know, in the flow of work. And that's, that's our long term ambition. And then the final bit is, you know, we're going to try and get to what we call the digital nudging concept. Yeah,
Chris Rainey 35:55
it's kind of like predictive at that point. Hey, Chris, you know, based on what I've seen in no calendar or less, you know, kind of nudging you it's super exciting, right? So, you know, we're seeing the evolution from agents being kind of order takers, as it were, to then being taking actions. And we've heard people say, you know, large action models, rather large language models. So, and that's kind of where we're moving to with Atlas as well, which I think it's really going to be the exciting part, yeah, as well. But thank you for bringing that to life for everyone about you know, and I'm assuming, in the future, also going to want to integrate some of the learning, the upskilling and re Skilling and learning pieces into that as well. Right? Quite naturally, you don't want to have another app that they go to for but that's another concern for people right now. I've got like, four or five different apps as well, so I think that's that's going to be really exciting to create personalized learning at scale, which you've never been able to do achieve before. And it's important, and I'm assuming, like any other agents, this can be multiple languages, etc, so many people that are at
Stuart Martin 37:02
yeah and natural language, natural language processing, etc, you know, so Yes is the short answer to all of that. I mean, you touched on a very interesting point. We are fully aware that our colleagues interact with over 80 different applications, you know, in a you know, in a month, some of them work related. The concept of an overwhelmed employee really does? You know, does is important to us, because an overwhelmed employee or colleague unproductive, they become stressed, there's a higher risk that they become disengaged in the work. And so we're trying to make sure we don't add to that by looking at this simplified concept. And so we have a bit of a mantra in the team and the work that we're doing, and that is, you know, we will make the simple we will automate the simple interactions. But for moments that matter, you need a high empathy, we will intervene and make sure you redirect it to a human for human interaction. Yeah,
Chris Rainey 37:57
that's important as well. And I'm sure there's part of that build with service now that you've included like, the why? The purpose, the mission, the culture, the values, into the agent, right? So I think that's important for anyone listening, that it has context. It knows who you are, it knows about the organization, knows your job, your country, etc, then it can be more personalized. So if you're thinking, Oh, why? I can do all of this with Jack GBT. You can't do all of that with Jack GBT. It's the context which means the difference to that individual to create a personalized Yeah. And that's where it gets really exciting when you can roll out your own agents. Have you heard of the you know? Have you seen kind of open AI's roadmap, where they talk about swarm, yep, yep. Right in all of that, yeah. So it's interesting now that it's almost like a battle amongst all of these companies that who's going to be sort of the master agent, as it were, because now we're going to have 20 different agents, and now can we connect them all to one master agent? So there's a sort of race in the in the, in this, in the landscape now, to become, you know, Microsoft obviously, is going to be one of those big ones that integrate everything has, I think they've got 80% of companies now, yeah, the big Fortune one companies use teams, yeah, so that's kind of, they kind of come up with better name than swarm submit.
Stuart Martin 39:18
It's a sort of doom and gloom battlefield. I know exactly, yeah, but, I mean, if you don't mind, I'll pick up on just one of that comments, and then also close out that particular one context is absolutely important. And hence the work that we're doing with the now assist tool, yes, which will give a very specific name that's got some content richness for our colleagues. So when we announce that they will link, they will relate it immediately. But context is absolutely important, and it's one of the reasons we've kept the, you know, our data, excuse me, internal. So when we talk about culture or values, not just a broader cultural comment, it's very much. We've got four, we've got four C's, care challenge, etc. So this will all come through, going through, especially
Chris Rainey 40:05
when you're doing that change transformation piece. This is a great tool to help lead that transformation journey that you unity, and we're going through right
Stuart Martin 40:13
now. And we can change the scripts, literally on the fly if we need to, if we change personnel, or we change, we tweak a bit. You know, some of one of our cultural aspects. We use Viva glint as our listening tool. So we do an annual survey, which most people do, but we also do monthly pop ups.
Chris Rainey 40:31
We have, will you try and surface those in the agent? Yeah. So
Stuart Martin 40:35
what I was about to say that? So what happened? We haven't done it yet, but we have the results of those surveys within 12 hours. Yeah, so either pause
Chris Rainey 40:44
there for everyone to take that in, because I remember the days of when I would give a call to a seer and they're like, Chris, I'm doing our annual survey. See you in three months. Just right? It's crazy. We can say that off hand like, you know, we've come a long way. We've come
Stuart Martin 40:59
a long way back to the earlier comments. This is about how work works change. We have a team that's focused on that. We use Vivian. There are multiple other products out there, obviously. But for example, we did an annual survey, which closed a few weeks ago. It closed at midnight with an 82% response rate, which was really good by the time our senior leaders came into the office the next morning, we were able to give sort of thematic content to them and say, Wow, this is, this is what we're seeing. And then we build that up over the next rest of the week. But within, you know, within 10 days maximum, we have detailed content analysis to feedback into the into the organization. And this is hypothetical, but if we saw a trend that we wanted to tweak in terms of culture or values or something, we can go into the tool and make that change Pretty much, yeah, within 24 hours, you
Chris Rainey 41:50
can also feed those insights back into the agent, yeah, and ask it for recommendations or insights and perspective, which is, which is so interesting that you could do that. Obviously you can do a lot of that in lot of that inkling as well, but the fact that you can put that within the agent and say, Hey, based on what I've just uploaded, and ask some questions to it like again, that would be such a heavy lift used to have to do. We've
Stuart Martin 42:16
also used other tools, copilot, Microsoft copilot, to take a look at the so I think we had 100 and 70s trying to get the number, I think 120,000 comments. Because not, you know, not all colleagues comment, but some colleagues have multiple comments for every question, you can add a comment. So that's how it kind of works. I think it's 120,000 plus comments. Through our survey, we were able to use co pilot to read through that and do a summation. And that's where you get the themes. It's obviously highly anonymizing, yeah, of course, the usual guard rails, but that's one of the, you know, fantastic uses we've been able to do. So do the analysis, close the survey, use, use some of our AI capability to do thematic search and push it into a visual report. Are you going
Chris Rainey 43:03
to integrate the agent into your teams?
Stuart Martin 43:09
Open debate right now, see where we go. I mean, to your other thing about swarms that was going to add up the latest and greatest from some of the big agents, obviously, the big service providers, is agentic AI, yeah, yeah. So, so exactly swarm in a nice and a bit more gentle name, we know we're probably gonna have to think about that, and probably mid next year on how do we then start leveraging agentic AI in a way that makes sense for us, but yeah, maybe
Chris Rainey 43:37
take a second just explain the difference for people.
Stuart Martin 43:41
So agentic AI is a place I'll skip the technology, technological architectural discussion, but essentially, when you have multiple virtual agents in your ecosystem, and I'll stick to HR, because that's the space I know, you have an agent responding to survey questions, you have an agent just doing general query management, etc, etc. So you have multiple agents. You then can have you can then build your own agent to do QA across those so they actually work. So one agent looking after another agent looking after another agent, the space where I think there's potential opportunity for us to to play would be in the payroll space. So you have AI, if you have a virtual agent helping you with running the payroll, an agent to help you check for discrepancies, a virtual agent to check that you are compliant with x, y and z, then you can have, from an agentic perspective, you can have an agent say, are they running smoothly? Are they doing the right checks? Do a bit of a Q and A where today you probably use human intervention for that, yes, but you could have an agent that kind of does the Q and A cross that process for all the other agents as well. And, of course, can also feed them work. That's the beauty of it. So can feed them work coming through. So a bit on the AI. Is, and I must be honest, but something we
Chris Rainey 45:02
playing around with that right now in Atlas. So we have Atlas copilot, which is almost, kind of think of that as your HR generalist, yeah, right. But then below that, we have a people analytics agent, a well being agent, a dei agent, an onboarding agent, and it's we're leveraging the agenda capabilities already, and it's incredible to see them interact and talk to each other, and you, and you're, in many cases, you're just interacting with Atlas, yeah, but it's then kind of distributing that work to those other agents. So you don't even know, because what we don't want is like a whole list of agents to interact with, yeah, but you can if you want to specifically talk to our people analytics agent. We worked with around about 20 VP to be valid, because they helped train that over a three month period and filled out the capabilities, even a prop library of specific questions, stuff like that. And we in the integrations. So we're integrating with a lot of the big companies that you just mentioned, actually, to bring in their capabilities in there. But it's incredible now to start to see the back and forth, and our agents are already conversational, yeah, so you're just kind of having a conversation back and forth. You know, still needs a bit of work. You know, it's like, it's not perfect, but we're on a smaller scale than you,
Stuart Martin 46:12
so we can but, you know, we learn from people like yourselves, who have that agility and probably a bigger risk appetite, who work in the regulated environment. But that's really, really good to know. And truthfully, I don't you know if you want to get the Gen AI embedded first and then work our way up there, but that's, that's where we're going. And if you don't mind me taking you back to us on one of your earlier questions, but what are the different skills and capabilities you'll need? Yeah, you know, in my world, that will be in HR, the stuff you just articulated, because I need to understand pay Wale, and I need to understand if I have a query around compensation, you know, it's not just a an IT question and not a tech. You know, don't need a technologist
Chris Rainey 46:49
very unique set of skills, that is, it's like a combination of, there does, there does need to be some expertise in the space, because otherwise, you know, there's no context. But then you also need very technical it's quite a Yeah, and that skill set, in many cases, like you mentioned, the prompt engineers, doesn't even exist. I mean, there's not enough talent, enough the skills back to your point earlier, the skill shortage, you have to, you have to start developing these, some of these skills internally, you know, like, if I, if I had to start my career all over again right now, probably being a prompt engineer. There's so I think that would be my first go to for anyone listening, if you want, yeah, but you know, because that is such a evolve just at the beginning. Now, talking about agentic AI, it's just
Stuart Martin 47:36
yeah, it's just like I said. I mean, it's exciting study is to be working in the workforce, because you can really, truly re imagine work. You know you don't have to go well, this is how we do it today. Therefore we need to do it slightly differently. I can sit back and look at that sort of integration of technology with work. So just
Chris Rainey 47:53
to very summarize that we spoke about a lot, what specific HR activities do you think will be replaced by AI because we spoke about, there's a marriage and some on some of those, right, with kind of leveraging, but which, some, which, what, some of the HR functions or activities you think will be eliminated, replaced?
Stuart Martin 48:16
Yes, these aren't HR specific, but I'll give you some examples that I see coming through. So I think the world knows in any function, but let's stick to HR, yeah, there's an army of people producing PowerPoint slides,
Speaker 1 48:31
you know. So that I still can't believe that. All right, okay, yeah, we still have small little armies producing
Stuart Martin 48:35
PowerPoint slides. I think that will fall away definitely. Some of the tasks where case summarization, or meeting summarization, where we still have already gone, that's that's gone, but we still have people doing that, but you do that, I think that will fall away. And I'll give you another example, where we do conduct cases, where we, you know, we do assessment of people when they've been arrested, or some particular colleague conduct issue. We now have the we now using case summarization in those conversations. And what that means is, I need fewer people in the room. But secondly, if I'm if I'm interviewing you for in a case context, I can really pay attention to you as individual. I can display more natural empathy because I'm listening and watching you, rather than sort of making physical notes. That's good point so that. So that's so that. So I fewer people in my context there, we start to see digital but learning bad disc bodies come through as well. So the conversations around you know, three of us need to have a workshop and then tell me what the learning content is, that kind of falls away, and I can go on and on. The other big area where we've seen some real pioneering stuff is what I call learning content development. Yeah, yeah. So you could literally between the two of us, within the next hour, we can create a high. Highly visual, very relevant, Context Sensitive learning program within the next hour with the tech, whereas that would have taken months and months and months, multiple iterations more,
Chris Rainey 50:12
and then you got to then you got to create in multiple languages. Yeah. That's one of our biggest focuses for next year. Is we're launching our learning pathways. Yeah. So Guillermo Miranda, my co founders, the former chief learning officer at IBM and Boeing, and when he when we first launched Atlas, he was like, the very first thing he was willing to solve for, the thing that gave him the biggest headache was we want to move away from static content to dynamic learning, yeah, customizable at scale. And now we've kind of built the foundation of the models inside of Atlas, and the AI kind of fills in the blocks, yeah. So now you can, within the second, I can say, I want up skill on x, y, z, and then AI can grab the different resources and content and create me the learning pathway. All right, specifically to Chris meets me where I'm at, I can say, I only want to do a video. I only want to do audio, right? And we've played around with this isn't live yet, but we're close that almost something was like a dream a few years ago, and it could be an hour, because actually make it this language, right? Yeah, it's and
Stuart Martin 51:13
if you take it to a large audience, yeah, maybe 100 people, you have immediate feedback, yes. So the end of that session, they go, this was this, you know, point five, it didn't do or the assessment. You've built in some online assessment capability, it'll come back to you immediately, and you go back into the studio, yeah, you know, make some tweaks, and then send it back out within, within, well, they can
Chris Rainey 51:35
even ask themselves. So we've, we have to be careful. There are some guard rails, but we've one of the features is that say, if you're on a particular module, when you feel like, you know, want to learn more, you can say, I want to learn more about this. And it will just add more content for that specific module and go a bit deeper for you, and it before you continue to the next, and then maybe the next. We've all gone these online courses where one of the six modules are very boring and isn't really, you know, skip that one. Double down more on this one, and then it shows you all of the different resources of where you're pulling the information from. So you're not just taking the course, but you also understand it. Okay, that came from here, that came from our service. Now it pulls that in to give me context, right as well. So all of a sudden, what would take you months of grabbing like those different resources, whether it's practitioner, benchmark data, whether it's you know, stuff from McKinsey and Gartner, your policies, you can wrap that all in one learning pathway that you normally be like talking about replacing people would be a manual lift from the team as well.
Stuart Martin 52:39
Absolutely. I mean, we can go to various spaces. The one others I speak to myself learning content designers is one, yeah, just chatted about, if in the recruitment space, you know, do I have to create these job descriptions for all these roles? Now I just gets done. I tweak it, make it content sensitive or content rich. So, you know, that sort of space is gone. How I post for roles, how I do assessment. I mean that the Yeah, so from the time I want to do workforce planning through the entire HR life cycle, any of those activities will be fundamentally changed. Argument here,
Chris Rainey 53:12
this is after we can keep talking forever. I feel like this was never going to be a half an hour conversation. Was it? We're kidding ourselves, and we thought we could get through this in a half an hour last question. First of all, thank you for coming in. It's been a pleasure to chat with you, and congratulations to you and a team away you got so far. That's no small feat, and I know that personally, because I speak to your peers every single day. So congratulations on the steps taken so far. And obviously we wish all the best with a roll out as we look to 2025 What are you most excited about?
Stuart Martin 53:44
So thanks for that, and I'll share it with the team. It's a team effort. 2025 you can be a pessimist and optimist. I tend to be an optimist by nature. So there's some other big ticket items happening across the pond, and I think that will, that will influence the geopolitical situation, but let's, let's bank that more specific to work. I think we're going to, you know, ramp up the Gen AI capabilities. Now, I fundamentally believe there's some huge opportunities that we will stumble across, and we be pleasantly surprised, as long as we open to it, as long as we do it with the right guard rails, with, you know, with the transparency that's needed in this space. We are mindful of bias, etc. But let's, I think we are smart enough as a human race to be aware of that. So I think the opportunities are infinite, and we will discover them as we go along next year. The other one I'm also very excited about, as I mentioned earlier, is, you know, the chance to really reimagine work that does that comes along sort of once, you know, once in a lifetime. Yeah. So really want to dig into that space and focus on that. The other thing I think, what I think we. Also come through next year is bit of work on what I call closing the digital divide. So we are, we are at risk of becoming almost digital elites in some spaces, leaving lots of people behind, where access is still a challenge. And so I'm really keen, and I'm hoping that we can spend a bit of time about closing their digital divide with the cross. You know, social mobility, whether it's from a gender perspective, whether it's from a access through geographic location, that's a space I think I'd like to play in more. And then we'll place, we'll see some more focus on that, because it's the right thing to do. But also, you know, the more you can tap into that capability to a larger part of the population. Again, the opportunities are, you know, phenomenal. So I think that that, that's going to be something that we're playing in 2025 in the technology space, you know, I think we'll continue to have a drum beat about Gen AI in other functions, and then a genetic AI, you know, everyone's going to try and sell that at the lowest common price. So let's take that, take that as we as we see it come through. And I think HR is again, at the, you know, at the front of, at the front of that race, in some cases. So I'm glad that we have a bit more of a pioneering role than than perhaps in previous 100%
Chris Rainey 56:20
I think it's incredible to see many HR leaders take the lead on this and be, you know, leading it, as opposed to, you know, the whole seat at the seat at the table. You know, if you're still having that conversation, I don't know what you're doing at this point, but listen, wrong conversation, the wrong conversation. Exciting times ahead. Super excited to see the energy and the passion you're bringing to it. Clearly you're in the right space as well, and I wish you and the team will the best until we next week. Thanks so much.
Stuart Martin 56:48
Thanks, Chris, thanks for having me. Bye.