 Welcome to the Agile People FIKAcast. We talk about how to navigate with agility in our organisations. Hi everybody and welcome to Agile People FIKA. We are meeting once a week where we talk about interesting topics and we are inviting anybody who would like to join. But today we have some Agile people here and we will today talk about agility in HR. How can HR use AI? I should have said AI in HR. And how somebody is joining us, it's David. Great, welcome David, we say. Hi there, David. Welcome to the Agile People FIKA. And then everybody can join. This is just an informal discussion. Today's topic will be AI for HR. What does it mean for HR that we now can use AI for all sorts of HR related stuff? And what will be the consequences? What are the risks? Things like that is something that we would like to discuss today. So maybe some of my teammates here would like to kick off with a question or a thought. Maybe Martin, would you like to start? Sure, I can try. AI is everywhere at the moment. It feels like I use AI a lot myself. I talk and interact with Shatgit GTP a lot to get some inspirational thoughts. So I'm sure any most things that we in our type of work do, needs and could make use of AI in different ways. So yeah, we can start with recruitment and talent acquisition maybe. That could be an HR perspective that could use AI and maybe help to streamline recruitment processes and even finding people maybe. I was thinking there are some risks associated with using AI for recruitment. What are risks that you can see? So one thing I remember seeing, it's a few years ago of course, Amazon, I think it was, they made an AI recruitment bot that was there to analyze and understand CVs. Of course, they fed the AI with data. It needs to learn based on some information and the data that I fed it with was all data that they had used to recruit people and when they did that, they also, as we do all humans, we have our biases and our behaviors in how we view different people. So that was also, of course, added to the AI and the way that AI behaves is that it sort of amplifies things. So what they got out from the result from that research at least was an AI that was super biased and not very good at seeing anyone except for white male individuals as a good candidate to any type of recruitment, which is of course wrong. But if you amplify something, that is what you get and I think that is a thing that we need to be aware of with AI in general. I know that the work has progressed of course since then so that you can reduce the risk of that but it's something that the AI engineers constantly have to work with in order to mitigate those type of biases and behaviors in the AI systems. I guess also that going through CVs is one thing and I mean there is some human interaction part of that as well to listen into behaviors that could be subtle but still important when you hire people. And of course an AI maybe can learn that as well but I'm just saying that there are things that are not factual or emotional that could be important as well. Yeah I don't, it will take some time but I saw that Tesla had sales AI which calls people who have been on the webpage put their phone number there and then it just calls that person and make the first call. That person is probably interested in buying a Tesla and then it talks to them asking them to come to a station where they can take a look at the car and you can find it on the internet. It's not super fast. You can see that there's a pause between when it's actually thinking about what it needs to say but it's super slick. Super real. Is there an expectation then and maybe just a wrong expectation that we think that AI is or we expect AI to be cleverer than it actually is. In that example that you gave there is you're asking people to look at CVs and you're asking them to say well you don't exactly tick every box of every row that we're looking for but let me be a bit clever about this and say well you know what I can see this job role and of course that's going to have those elements in it so you're asking it to exceed that capability. If I think about some of my career history then with implementing technology you always get that expectation that it's greater than it is. So for instance if we have a move to the cloud in terms of technology then I've heard clients say oh well it's just a whole load of little robots just running around processing these transactions end to end and I don't really need to do it. All I need to do is just oversee it, align it and the end result's produced and I didn't really have to do much around that so there was an expectation that only jump in if something goes wrong but hey the robots will take care of it because they're clever enough to do it. There I really think HR plays an important role because if you only leave this to the engineers you will not get to say because you need to involve people who have understanding and knowledge about the human behaviors and one thing I can see because today in some recruitment processes you use like personality test or IQ test or these kind of tests so if you don't say yes or if you don't qualify on this you go to the good buy box and I would say many of these candidates could still be a candidate for the possession so I think it's how you design the selection phase because with AI you can do so many more parameters which you don't have in your head today and you can also if you do the right design you can also eliminate some biases we have for example I think age discrimination ethnicity could be one religious could be another one which you will take more away because people might have some wear some things on their head which can be discriminating in some aspects for someone one you don't see this with AI AI I think that's a good thinking I was thinking my continuation of that thought was that recruitment does not need to be that reductionistic I mean we use IQ tests and other tests in order to it's a reductionistic way to make the recruitment process more efficient but when you make things efficient there's also risk of missing things that's important and AI could interview people potentially as a first input and gather the data and make it easier for the human to make informed decisions instead of just using IQ tests but also another thing is what you say but also today we do a lot of recruitment based on what you have been doing in your former positions it should be a little bit this is the position we have what would you like to do if you got the job and you could start describing how you could see the organization how you could see the job how you could see whatever that is important depending on the job yeah instead of only basing it on a past experience or a good addition what would you like to do for the future and how would you like to contribute to our vision, our purpose, our goals and what can you bring to the table here I guess it comes down to how we want to use AI and which decisions we leave to ourselves and what the support is we're asking for yeah I mean I see AI very much in its infancy in that respect that it's not trying to take over your job and its entirety it's taking over the heavy load or the standardized type processing and so leaves that skill to the human being who actually operates it whether it's in the design phase or whether it's in the output that comes it's really going to be your decision you're not asking AI to be a decider for you it's merely a naveler and I think it's the good thing that at this stage you're not really going to have that abuses or real requirement for regulation because you're still in control of it it's not that technology is driving you it's the other way around you're still in control of what you need it to do there's a conference actually happening today I don't know where it is and they're talking about that very thing listen I know we're all talking about AI I know we see it as a threat to terrorist threat or other abuses that could come but it's about calming down we're still very much in control of it and we use it for what we need to use it for we'll get it wrong sometimes but hey that's experimentation we just keep working with it it's there as an enabler they used to be a role called computer that was a person that was yes absolutely we don't have that anymore but people are still operating and using computers to do their work and I use a calculator sometimes I don't do all my math in my head it's very good to have a calculator actually even though it's in my fault nowadays yeah but then you know how to use it what I fear a little bit with HR systems that will come is that you use your traditional way and you take your work day or success factor process and you build it into IE instead of starting with from scratch and think how can we do this differently and I saw that recently as well I was at a work human conference back earlier in the year and they did exactly that their AI tool was on performance reviews they started off with a blank sheet of paper for everyone and it was literally can I get AI to write me a bog standard performance review make it professional, make it funny make it serious, make it whatever and then it generates the text and for you as the human then go on and then you personalise it so it's basically just taking that blank sheet of paper generative text to say oh well I like that, that's pretty good but let me just reshape it in terms of the role that we're actually looking at and the way that that person performs in the role you're still using that human skill of personalisation because you know sometimes when you try to write a performance review and we've all done it whereas you're having to write from scratch and you're ripping up the paper and throwing it away and scribbling more and all that sort of stuff you know, generative text is just doing exactly that and then your manipulation there afterwards is intended I guess to be fairly minimal Yeah, I use jetki to be as a sounding board I just bombard it with my thoughts a lot of different things different perspectives and I ask it to okay, can you write something about this and then I get a very nice and tidy, I don't always agree with what I see but it makes some sort of order of the things that I have in my head with a very competent way to, much faster than if I need to write it down myself I could have done it but it's much faster doing it together with chat UTP Yeah Okay, so we move to a topic outside of recruitment for HR I have a list of things that I would like to show and see if there is some topics that you would like to discuss more in detail there are some examples on this picture that I'm showing now how you could use AI for HR So is there any specific box that you would like to comment on or deep dive more in of this specific because when you ask chat GPT about in what areas will I change HR I will just mention all areas but this is a little bit more specific it's not the performance management or learning and development or talent management or employee engagement but it's more you could do this specific thing and I like that a bit better Could I ask Pamery because I'm looking at a number of these and they're all kind of data driven of a fashion because the data exists and then you could easily pair what you need to do to drive that out if I was looking at the one that jumped out with me immediately was nurturing the company's culture just because I don't believe that that would be data based and therefore easy to determine it's probably based on data that we are feeding it with what are the values that we would like people to do more of and so on and then it will send out little nuggets or messages maybe or I don't know how it works but you could imagine that we are feeding it the culture that we want and then we ask it to what are your suggestions for how we could work with that for example and increase those values I'm thinking that many of these let's say the analyzing the state of employee engagement for instance as well it's and I see it in others as well it's a possibility I guess to get AI to to provide or to measure the health of the system we're working in so what is the flow of the company and what are the behaviors and map that to culture maybe at the end so if someone is keeping track of how all the delivery teams for instance if you're in IT set up so how are they performing how are people acting in the system how much wait time do we have for decisions you could get the good overview of how it's functioning I guess and then if you could connect that to cultural maybe you can draw conclusions about that and try to nudge the system in another direction or highlight things what the next for instance could be one thing good also you can use it like we said you I use it to create text you can create personalized information that's more attuned to specific people connected to the company culture that is more appealing to them even though it's the same thing but it's more directly aimed to a specific audience that's something you can get help from from AI I think I don't know if it's super connected here but maybe you can even do leadership coaching or leadership trainings using AI I've seen a different type of coaching tools that is using AI in order to train people to be better coaches for instance they are starting to pop up to pop up and that could also of course nurture the company's culture I actually asked JPT and they said it can facilitate communication both employee engagement personalize learning and development diversity and coaching employee well-being it's a lot of these things fostering innovation one of the things I can see is really the first one so facilitating communication because there you can do a lot and personalize it a little bit to get an individual and also you know company new processes, new values best practices and you can feed a lot of information and they tailor the messages to an department a whole company or individual so what about our current HR platforms or HR IT systems what do you think what is the current situation for them, do they keep up with the pace here or do they still need to be more mature before we start thinking about buying new IT platforms for HR. Thank you start to see some good examples of where AI has been used in some systems I think the larger ones will try and adapt not sure how they will succeed, they need to think a little bit new I think but it's a lot of knowledge also in these companies that you can use I presume but I think that in the future we will see new companies that are not there today or not known at least they might have started there. Yeah maybe building AI into the platform from the beginning and then of course it's much easier to make it work the way you want to work I think also you can use this because not everyone wants to be employed some people want to be freelancers so I think you can use this also to build in when to use freelance competence and when to use internal companies I think it will be more we will not have these standard roles in the future it will be more that we pull the competence or the skills when we need it So in the HR system that I've seen is never really fit for purpose in terms of a wholly integrated solution for the many components of HR and so as much as they try to do it what it sounds like is that either AI is going to help improve streamline that standardisation within a one product or it's still going to reside as a best of breed applications and so therefore with the fluidity or the unknowns of HR type role in the future maybe that's the way to go and AI sits above that and orchestrates the better ways of working I think you already start to see that you have systems that are good at performance or recruitment and that will probably accelerate more. I think these large work success I think they would struggle a little bit because of the pricing structure and they need to do everything So for many years there has been a consolidation in the HR system market that maybe it will now go the other way is that we will see more best of breed solutions powered by AI and that's a probable outcome and because of AI I think that is also possible now because it can grasp several systems together which was a little bit harder before because then you need to have good data warehouse to make that happen. The integration will probably be easier also to transfer data between different platforms There's been a lot of talk about the fourth industrial revolution and AI has always been one of those items that people talk about that is part of the fourth industrial revolution so I think we are only in the beginning seeing that so there will be a lot of disruptive new innovative tools and ideas connected to AI also in HR so I don't think we can foresee what will happen actually as usual as usual I guess we see it more as an opportunity than a disruption but it will be destructive it will be making traditional HR tools and I'm thinking about they will still prevail probably most likely but it's compared with Airbnb for instance and the hotel business when Airbnb managed to get into the business it was quite destructive for the whole hotel business of course they are still there and they've adopted adapted not all of them so it made other things change and I think AI will do the same so summarizing the discussion we think we will get rid of boring routine tasks pretty quickly but for the more sophisticated AI use or solutions we will probably need to wait some time and the platform for HR the market there is somewhat immature still but I would say whether we like it or not AI is here to stay and if you are in business you need to relate to it you need to adapt to it you need to understand how to use it and you can then as always as we do in agility we run small safe to fail experiments and then we learn and then we improve and this is the way we need to handle AI as well it might take and we need to follow the development of the new HR tools that are coming now because there will be applications probably that are being born that are appearing now a lot will happen in the market during the coming years but I think the most important thing is that we should have fun with it we should try different things have fun and play and see how can we get rid of the boring tasks first with the help of AI in HR I agree to that good summary so when do we see each other next time Wednesday next week we have a topic yet we were talking about discussion team development but we will put it on the LinkedIn feed as normal okay yes so next week same time Wednesday 3.30 CET maybe another time where you are we will see you again with that thank you for today thank you all of us who have been here today take care everybody take care