Human-Centered AI: Prioritizing Human Connection for Greater Impact

So, hey everybody again, really excited to be here with you. And I spend my days as a researcher, as a, as a leader, trying to bring data, bring insights back to all of you, to be more effective in your work, to hit the, the goals you have to, to really solve the big problems your organization is facing and your people are facing as well. And so I'm gonna share for a little bit about what we're seeing on the, the forefront of ai. We're gonna have some different talk tracks today. I get the, the joy and the opportunity to talk about AI to start off. And what I love about this topic, honestly, is some of you might hear that and you're like, this is an, this is an HR session, right? This is about people, this is about humans and connection. And why is there an AI piece on this? And I'll tell you why. When I was an HR leader, when I was running an HR team, when I felt that, you know, sweat on the back of my neck, 'cause we've gotta get payroll done Friday and we're, we've gotta do this investigation and I've gotta re look at these requisitions and all the things you have to get accomplished. Like that list of, to-dos is always getting longer every single day. And I know that feeling, I know that sense. But I also know, based on the research we're doing, based on the technologies and tools that are out there, we're seeing a lot of leaders just like you. They're saying, I'm gonna use some of these AI tools to help me be more impactful, be more human, be more connected to the work and the people that I'm supporting. And so what excites me is employers out there are looking at how do we use this in, in hiring? How do we use this in employee development and growth? How do we help each of our individual people understand what they're facing, what their challenges are, but help them figure out a career path for them. And AI is being used in all those ways. We'll hear about that in a little bit. We'll get into the panel conversation, we'll talk with these leaders about things like retention, about coaching our managers, some really specific kinds of ways that this can be used. But I want you to hear that I am not a fan of AI for the sake of keeping your people at arm's length. It's not about saying, let's use this instead of connecting. No, it's, let's use this to make that human connection even deeper. Right? I bet when you are meeting with people now for the first time, be like, I'm gonna go scour their LinkedIn profile and try to learn more about them before we meet up, right? That's a common thing. It's the same thing with ai. It's let's make this connection, this engagement, a little bit more rich, a little bit more deep when we do have it. So that said, I want you to know that one of the things we see in the data, I'm gonna share a couple of different research points with you, some different resources here with you. One of the things that we see in the data is that there are some very specific things that AI's good at and some others that it's not so much, it's not wired for, it's not meant for these things. And the, the things that AI is really good at, actually put out a report recently that talked on some of the tasks and the jobs and things that AI can and can't do. And I'll make, uh, we'll make sure to get that one into the links here and share that in the chat so you can bookmark that, come back, read it later on. It really looks at the things that you're good at though as an HR leader, the things that probably drew you into this in the first place. The good news is that AI can't do many of those things very well. It can't help you build a great culture. It can't reach out and intentionally and personally engage every one of your employees. It can't help you to build this, this, um, bench strength of leaders as you're getting ready for succession and growing your organization. And we need leaders as a fuel for growth. It doesn't do all those things. And those are things that probably get you excited, things that you drew you into this in the first place. However, the things that we don't all love that much. Like looking for payroll errors or trying to read through a spreadsheet with 500 employee open comment responses to a survey. Those are hard. And those things are difficult for us to do. And AI's really well served for those kind of purposes. 'cause it can do the analysis, it can do the preliminary sort of fact finding and checking for us. And then we step in as the humans, as the experts to actually solve whatever those problems happen to be. So that's one of the big things I want you to hear is there, there's a clear divide and we'll hear more about that in the panel a little bit, what it can and can't do. But also the decision we have to make. We as HR leaders are at the forefront for many organizations right now saying not just what can it, can it do, right? But what should or shouldn't we be using this for? Should we use it for this purpose or not? Is there an ethical reason? Is it responsible? Is it biased? Can we use this and solve problems or is it gonna create other issues down the line? And speaking of bias, one of the things that we see in the research, and one of the things I speak on this a lot and I hear consistently is the question comes up, Hey, yeah, I read this article, I read this headline. What about bias? How does ai, you know, solve for that? Is AI creating and causing this? And I'll tell you the honest truth right here. Many times when those headlines pop out and say, look, the AI's biased, what happened is a company was already biased before that. Maybe a manager, you, you may have one of these managers. It's like, oh yeah, we can't pick, uh, that kind of person. 'cause they're not a, they're not a fit. They're, even though they can't tell you why, they can't give you something more clear. They just say, oh, they're not really a fit. But that manager never picks that kind of person. Well, that creates a pattern. And when you see the news about ai, it's looking at those patterns and just amplifying them. So we didn't create bias. It didn't introduce that one. It didn't exist before. It's actually using something that's bias right there in the moment. So again, one of the other resources, I'll make sure we drop in there, we did a piece a while back on how AI can actually positively support DEI, um, employee belonging, those kinds of things. Because all the things that you've see in the news are like, it's bad. It hurts that it hampers that and holds us back. And in fact, it actually can help to level the playing field. It can make it more fair. It can help us to make sure that we don't treat women differently than we treat men, or we treat people of color differently than we treat our other employees. We don't have to do that. And AI can help us get better at it. And then finally, the last thing here, I wanna like the big point I wanna make for you before we transition and bring our panel in just a few minutes. The last thing I wanna bring to you here is the perspective of this. It is entirely possible to go too far. It's entirely possible with automation, with AI to go too far and to automate too much and to, for us to rely on these things too much. I I'm like the, the big proponent and advocate for this because I think it's gonna get us closer to what I mentioned earlier, the things that drew us into this. The things that bring out our best human work. I think it's AI's gonna help us take care of those things on our to-do list to get us fo focused on that work, which is great, but we've always gotta be careful. We always been cautious of where to draw the line of not going too far. So there was actually a really interesting study someone sent me recently and I I soaked this whole thing up 'cause it's so fascinating if you can't tell already, I'm a research nerd. Okay? So there's, I have that sort of behind that makes it more exciting. But the other piece I'm gonna make sure and share here in the links, um, the resources is a piece of research from Harvard. They actually went through a group of recruiters and they gave them what they called strong AI or weak ai. And so the strong AI was, seemed very predictive, seemed very competent, seemed pretty broadly applicable and helpful. The weak AI was not as predictive. It could only do a very narrowly defined thing. It couldn't say, okay, let's look at all this and pick the best candidates. It could just look at one specific piece of the hiring puzzle. And what they found was that recruiters who relied on that strong AI began to take their hands off the wheel and just let it go. And ultimately it made worse decisions. Ultimately the choices they made for who to hire were worse than those who used what they called weak ai. Something that required the human to stay engaged, keep your hands on the wheel, to stay in charge of the decision all the way through to the end. So that's my encouragement to you that it's, if you wanna put your head in the sand or if you wanna, if you're like, I just mute this and I'll come back when they're talking about the people stuff, I'm warning you, you're gonna miss out on the chance to really emphasize and focus on and focus, deliver better people outcomes. Because the tools can help us to be smarter. Right? Now think about this. If I give you a random employee in your organization, any random employee, could you tell me the two or three things that are challenging them that are holding them back that are issues they have with their manager? Probably not. We don't have the bandwidth to do that for all of our people. But if we're using tools that help us to see those issues, see those challenges, see their frustrations, see their hidden skills that we can't even see ourselves and that allows us to then show up and engage them and deliver better results and make sure their managers serving and supporting those development needs they have, suddenly we have more of an impact on the people. And that excites me on so many levels. Alright, so I'm, ooh, I'm gonna like hop off the the soapbox for a second. Take a deep breath because I actually have an incredible panel here with me to talk about some of these things to really dive into the conversation and talk about some, some key parts of what they're seeing, what they're excited about, the stuff that, uh, the kind of their perspectives on the bigger AI conversation overall. And so I have, um, Zach Frank here with me from Freeman Company. He's people analytics, AI expert, super sharp. I have Amit here. He's, uh, from general assembly, HR transformation, digital transformation, like big picture HR strategy, that's where he lives and he can eat, breathe, sleep, that stuff. And then Brad Wilson from Perceptyx, his focus background in IO psychology, and again, we'll let them do their little intros in a minute, but I was excited and I'm gonna ask him some hard questions about the human versus the, the technology in this and where we sort of balance those things out. So we wanna transition here, bring those awesome panelists up with me. I would love to. Alright, perfect. I'll see all your pictures. We're all there. Great. So welcome to all of you. So glad you're here and hopefully I set the stage with some stuff. Maybe I set something controversial you wanna pick apart, which would be great. I'd love that. Alright, so really quickly before we get into the, the meat of the, the content and start diving into the, the conversations about, um, some of the key things that you're seeing and, and all that good stuff. I wanna do a really quick round of intros for you because I gave your title, but that's about it. So Brad, you're my first one on my little round robin here. Give us a quick about Brad, what you do, who you're, Yeah. Um, great to be here. Appreciate the opportunity to share. So, uh, I've been with Perceptyx almost 14 years now as employee number nine when I was hired. We're quickly approaching 400 people now, so it's been quite a, quite a ride. Uh, background. I have an MBA and also a PhD in Iowa psychology. Um, also a husband dad to four amazing kids. And, uh, currently living in, in southern California. Alright, another data four. Okay, that's me too. Awesome. Perfect. Alright, next up on my list, my, my friend from across the pond. Uh, Amit, tell us your story, who we are and what you do. Yes, uh, please. And excited to be here. My name is Amit. I have 18 years of experience in regional and global roles across Europe, Asia, and North America. I've predominantly worked in pharma consulting, financial services and technology industry. I currently head up global HR transformation at General assembly. And, uh, pleased and excited to learn from each one of you. Alright, wonderful. Glad to have you here, my friends. And last but not least, Zach. Frank, give us your quick intro, who you're and what you do, please. Yeah, uh, Zach, uh, I've been doing different types of data and analytics for about 10 years. Started out in logistics and then did healthcare for a little while. And now I've been doing people analytics for about six years. Uh, and I've got a wife and two kids and live in Nashville, Tennessee, Nashville, Tennessee, just a couple hours north of me. Awesome. Wonderful. Well, glad to, to have all of you here. And again, as I told the audience, uh, all of you a few minutes ago, each of them brings a different set of perspectives and I think that's gonna make this a, a richer and more valuable conversation for all of you. I see some Q and a's coming in. Select that, by the way, drop those in the q and a. I don't know, we'll have how much time, we'll have to get to all that in the conversation right now. But we'll be capturing all those so we make sure everybody gets an answer. If you have something that you're curious about, make sure to drop that in. We wanna address that if we have some time. Okay. So let's dive into the good stuff. So Brad, we'll start with you. As I mentioned a minute ago, like your background in examining human behavior as an IO psychologist, really looking at decision making, all those kind of things, it's easy for me to imagine that someone with your background might think, well humans, yes, great ai, maybe we don't need to be using that in the workplace. Yep. But I'd love to hear from you, what's your take on that? Is that true or not? And to what extent? Yeah, so, um, I I really appreciated your intro and kind of the way that you position this, but, um, you know, the question kind of implies that there's this conflict between people and, and ai. And I think the reality is there is a lot that both sides can contribute to the other. So at perceptyx, we are using AI currently to ask and answer research questions that were previously unanswerable. Um, and, and an example of that, uh, there was a client about eight years ago, we were looking at DEI, this was a tech company that headquartered up in the Bay Area. And we did an analysis of, of really inclusion, did they have an open, inclusive culture? And we looked at similarities and differences between when men and women by job level. And at a certain point in the organization, there's this clear divergence where, uh, scores for men went up significantly in each senior level, and that wasn't the case for women. And so once you get into the director level and above, there's this significant gap. And an executive asked a great question in the meeting. They said, is this evidence that men and women are being treated differently? Or is the problem that they're being treated the same, but the needs and expectations are different? And based on survey data alone, we couldn't answer that question at the time. But now when we pair AI with things like passive data collection and we're able to look at things like responsiveness and positivity, we can actually get into some of the nuance of those situations. And in some cases, what we're finding is that there is a difference in experience and behavior and in other cases is the difference of perception. But AI is actually enabling us to delve deeper into solving some of those tricky problems where, you know, survey data alone can be more subjective. I would also say there's an opportunity for AI to benefit from what we've learned on the behavioral sciences side, things like validity and reliability. Does the data truly measure what it is that we're trying to study and manage? Um, that's something we need to know. Um, you know, and, and how do you formulate good research questions and a research plan? You know, simply throwing more data. And especially when organizations talk about structure and unstructured data, the, the bias seems to be just more data, more data, more data. But we also need to know how do we ask and approach these with the right questions. Um, so that's really where I do see both sides benefiting one another and really more an opportunity for partnership rather than competition or, or conflict. Excellent. I really like, I like the, uh, the piece on like, we're now able to see kind of underlying factors here because guess what? There's still a job for us. We've gotta step in and actually do something about it. We can't just say, oh, that's interesting, now let's go back to the next problem. No, it's, now that we know more about the root cause, let's step in and actually address that policy systems, right? All those kinds of things that we can do and culture, this piece of that as well. Yep. Yep. Awesome. Wonderful. Okay. Zach, you're next on the list. So you, I see you as, as an expert and like what's possible with ai, right? You're, you're focused on people analytics and data and you probably were like your heart flutter as, as Brad was sharing on, on some of the data stuff that you can find out with AI a minute ago. Um, I want you to tell us, talk to us for a second about what is possible, where's that line of what we can and can't do or what should or shouldn't do. I know you have some interesting like, thoughts on this and I think it's a great way for the audience to wrap their head around this bigger conversation of, of AI and what it can do, but also what we should be using it for as well. Yeah, and I think those two things, I mean, the first part of it, I guess I would say is those two things really go hand in hand. So as we think about, you know, from an ethical or a philosophical standpoint, what should we be doing to get the most leverage outta that? We really need to understand some of the technical limitations so that we d don't accidentally go somewhere. We don't mean to go. Um, now if I'm gonna like grossly oversimplify, all the AI that's out there on the horizon right now basically falls into a couple categories. You've got systems that take data that they've seen before and they take something new and they say, is that a like, or is that different? And they just sort it in that a like different, a like different, right? Uh, then you've got sets like Bard or like chat GPT the take in what they've seen before and they synthesize the appearance of something new. Uh, and then there's, uh, this other class that I love the most that it's fallen into this AI category of just the automated application of business rules, right? This has been going on for a really long time, but now the AI monikers so hot. So a lot of stuff gets thrown under that. And they've got hybrids and hybrids can range all the way from like Boston dynamic robots that are tying a lot of these things together. But fundamentally they're just doing the same things, you know, they're kind of chaining them together and whatnot. Uh, and then you've got other hybrid systems that, that kind of fall under that, right? Um, you know, think of autonomous vehicles or even the wrappers around like chat, GPT, the di, the distinction between its underlying model and the website that you interact with, where you've got a little bit of, of hybridization going on there. So as, as you understand those things more, you get a better view of those technical limits, right? Because you don't want a word predictor deciding who you should hire. Uh, and that's, you know, Bard and chat GBT. If you take away the wrappers, really what they're doing is, and again, gross oversimplification, they're looking at a list of all the words in the English language and they're giving a score for how likely it is that that word comes next in the sentence that it's righting. And then it writes that right now, when I describe it that way, no one has any thought that it should be choosing who they hire or deciding how their program should run, but we kind of give ourselves over to it when we don't think about how those things work. Uh, and just to tie in one more example of how this can get us off the rails, kind of going back to bias. There was, um, a, a hiring assessment vendor that, uh, we were looking at a couple years ago. And, you know, I think hiring assessment AI can be really helpful, but one of the things that we asked 'em was about, well, how do you monitor for bias to be sure that the system isn't behaving in a biased manner? They said, oh, well, we don't ingest any data about demographics, so the system can't be biased, which is maybe the worst possible answer that they could have given besides we're being intentionally, you know, racist or sexist or whatever. Because all it means is they're not watching for the things that it's doing. If all it's doing is taking data that it's been fed before, if your organization has been biased before, that's all it's gonna spit out into the future. That's excellent. I wanna come back in a little, probably at some point. I've gotta look through my questions, but I wanna touch on, you mentioned like ethical, and I think that touched on a little bit there. Like, we have a responsibility as the HR leaders, we were biased experts yesterday, we're watching for it and aware of it. We can't just say, well, the AI has probably got that care taken care of and we'll just move on to the next thing because we still have to be the bias, bias experts in the business asking questions like the one you just just laid out there and having the wherewithal to say, wait a minute, hold on. That that doesn't sound right. Because we may not think about someone's skin color or their gender or their age when we're making decisions in the company, but we look at the outcomes at the back end of that and say, wait a minute, this is actually harming this population. What went wrong? What do we have to do to fix that? And that's still a burden on us. It, we can't just abdicate at the end of the day. Okay. Right. Great. Well, and you know, the, the, it's women tend to understate their accomplishments and their certifications on their resumes. So if we just say, okay, well we're looking at it a base like we're not taking into account whether they're male or female, that kind of goes back to what Brad was saying about well, treating them equally. Treating, treating the resumes equally, may not be treating the people equally. Hmm. Yeah. Okay. Uh, uh, awesome. So, um, Amit, I wanna come to you, I wanna talk about something practical here. So we heard some, some good things already. One of the things that I know you are really curious about, and probably the people listening in here, one of their things that they are responsible for is retaining employees. So when you talk about how AI can predict retention of employees, how that fits into this puzzle and how it can help us maybe to help, uh, mitigate some of the risks there of people leaving the business, we're using some AI for that. Yeah, I think, uh, you ask a very valid question, and I think, uh, you, you talked about business outcomes, foreign organization retention clearly becomes a number one agenda. But even if you, uh, step back a little bit, a lot of researchers across the globe, and this is not new, but in the last five or six years, uh, career advancement opportunities has become the single most important driver for a number of individual. And so is purpose led work. So what AI may do, may not do is a different ball game altogether, but I think from an HR standpoint, from an organizational standpoint, business leaders are constantly interested in knowing how their teams are doing. What is their morale looking like? What are some of the shifts that are happening at workplace? What are the changes in their lifestyle based on who they are, what stage of life they are? And as a result, they're constantly looking for data. Now, where AI may come in is an opportunity to, to look at large number of data sentiments across a range of employee life cycle or journeys within an employee life cycle, and then typically look at people who are at the greatest risk of being lost to the organi, uh, to, to the external environment. And as a result, a number of AI led platform allows business leaders and HR leaders to look at that opportunities to see who are these people, what sort of demographics they fall under, what are their needs? So this allows an HR leader to look at specific insights, specific data set to then plan for an intervention to look at if there are people, practices, processes that need to be looked at differently. If there are career advancement opportunities, if there are needs that an individual or a demographic profile may differ. And as a result there lies the greatest opportunity. We also know that a lot of tech organizations have found it, uh, difficult to retain talent and throwing dollar at them is not not gonna solve the retention issue. Mm-Hmm. And that's why, uh, cleaner data, insightful data is what a lot of business leaders are after. And what AI platform provides them is an opportunity to look at the data differently and not blindly looking and trusting the data, but unearthing that conversation and opportunity to have a conversation with those individuals. Yep. Excellent. Brad's over there like nodding vigorously, like Bobblehead, uh, what's your thoughts, Brad? I know you're some, Yeah, so I, you know, I think from a, um, you know, if you look back at, um, production, um, you know, mass production, there used to be a direct association between quantity and ability to, to customize. So you had these small boutique organizations that were making, you know, a dozen widgets a year, but they were highly customized or you had, um, you know, mass production, lots of volume, but everything had to be the same. And then as technology evolved and advanced supply chain, you know, changed in the eighties, nineties and into early two thousands, there emerged this concept of mass customization. How can we mass produce higher volume, but also give the consumer more choice in the process? And that's really what AI is enabling in terms of people practices. It's saying, okay, we can support an organization at scale, but we can also understand and address their unique needs and situation in a way that we couldn't do this before. And so, you know, that's the interesting thing is, you know, you talk about going too far and you know, it becoming dehumanizing. The reality is when it's done right, it actually enables the organization to create a sense of connection. And one of the, the concepts that's critical to this is the idea of benevolence and belief in the individual's point that the organization has their best interest in mind. And I think that's an important concept because when we do have the individual's best interest in mind, then we're able to offer personalized recommendations, solutions, and then also gather their feedback to say, you know, is this working or not? And then constantly learn and evolve our, our people practices using ai. Um, you know, I I think that's really where the, the magic comes together, you know, in, in this area. Alright. And Ahmed, I like I saw B Brad like bouncing up and down in his chair to, to share that, but I love, I love the point there and you got some really good kudos from the audience there. Clean data, valid data, right? We ha we are making some assumptions that we're using some of those things and to make the predictions make the recommendations. And if we don't have those as sort of the foundational pieces as you mentioned, then we're, we're gonna make predictions in the wrong direction. We'll make the wrong decision or wrong prediction even faster, which is not what AI was intended to do in the first place. So I'm, I'm, uh, definitely got some good kudos from the audience on that one there. Um, Brad, I wanna come back to you. So your team has actually worked on using AI for one of the biggest challenges for US HR leaders. It's actually such a big challenge that we have a separate track today just talking about supporting managers. And I would love for you to take a minute or two and talk about how the AI fits into that puzzle. 'cause like a lot of HR leaders listening in here may be picturing that that challenging manager, we'll just say that they're kind of like bumping heads with and things. And um, they may picture as well that this the range of leaders they have some of them that are really trying to serve their people, they just don't know how or don't have the time to get to them. Just talk about how that works, how that's actually supporting them and what it looks like kind of in a day-to-day practical sense. Yeah. Um, great question. So one of the common criticisms that we hear, and obviously, you know, perceptyx, our, our core focus is employee experience. Uh, we've been doing, you know, large sense and engagement surveys for, for over two decades now. Um, one of the common criticisms or complaints that we hear from managers is, you know, an organization does a survey and then it feels like they get this whole laundry list of to-do's coming out of the survey and it's all these things they need to fix. And it's more stuff on a plate that's already very full and it feels like something's gonna gotta give. So, you know, I think when it comes to the analysis and it comes to the integration of ai, um, you know, we're all hearing about doing more with less. The reality is this creates opportunity for managers to actually do less with greater impact. And that's what an AI is enabling. So if you think about like the, the 80 20 rule, 80% of our, our outcome comes from 20% of our effort. You know, if, if AI can help us identify that 20% and enable us to be more effective, that at that while also automating streamline and, and reducing some of the waste discrepancy and strain around this low value added activity that still needs to get done, it really can solve, you know, both purposes. The other thing that we're finding and, and Perceptyx recently earlier this year acquired humu, which is the organization that that's been focused on, on nudges. Now we are in the process of integrating those nudges. So they're informed by survey people, analytics and HRIS data, but it's actually feeding those, you know, small insights and recommendations to users and then giving them the opportunity to provide feedback. So at perceptyx we've had something we call actionable insights for, you know, almost from our inception. And one part of the actionable insights is, okay, if, if this comes up on the survey, here's some recommendations. But those recommendations were, were fairly broad because they were written in a way that they would work in almost any, you know, circumstance. But now what we're also able to do is to gather feedback and say, okay, in this situation, this recommendation didn't work. And the reality is it's not, doesn't mean it was a bad recommendation, but something about the context didn't work for that individual in that situation with that recommendation. And so the system is actually able to learn from that and say, okay, you know, here's this response profile or user profile or persona. We're saying it, it's not getting traction here, but it is over there. And so over time the system is getting, you know, smarter at that. So again, it's alleviating some of the pressure where managers often feel like surveys result in all of this extra work. And instead what we're doing is we're saying, we're not telling you to do more. We're actually telling you to, you know, focus here and here are some tips and insights and small things that you can do to really move the needle linked to some of those business outcomes. Like, like Amit mentioned of, of retention, of performance, you know, whatever the, the managerial dilemma is within that organization or business. Um, so, so it's really bringing about positive behavior change and we're able to associate that back to better outcomes for our customers that we're partnering with. So there's, there's a lot going on there, there, um, but it's exciting to see how it all comes together. I, I'm seeing a tie between that and what, uh, Mette shared a minute ago about the, the retention piece of predicting that, because if I as a manager can do one of 10 things and I know doing this one is gonna affect support positively six of my people, but doing, this one's gonna just serve one person. They're like, I've gotta figure out the highest leverage activities, like you were saying, not just doing more things. Because putting more on their plate is always a challenge. It's always gonna be a battle. It's always gonna be a, a struggle to do that 'cause they're just a finite time. Mm-Hmm. We said, Hey, instead of just trying to do all those things, if you do these three right here, it's gonna have the, this kind of impact potentially based on what the data are that we see that's so, so thrilling. Um, that actually connects back to, uh, as well. So I'll give a quick shout out. I interviewed Zach on, uh, on a podcast. Goodness, it's been over a year ago now. And one of the big takeaways from that conversation was, we as HR leaders, and you're hearing it from all our panelists to your audience, we have so much data at our fingertips. And when the organization is trying to make a decision, they don't need us to say, well, here's all the information, here's all the data. Zach's big stance in his sort of platform is we need to use the data to give them a recommendation that we can stand on, like, use our credible expertise and tell them what to do because they brought us in to do that job. And I think that's a, that's a another thing that ties into this. Like the nudges deciding what to do next, like that sort of thing. Whether you have the technology or not, you have, you have the data on hand to make that recommendation. So, okay, wonderful. Goodness, I'm like nerding out over here in the best possible way. Okay, Zach, you talked a little bit about, you mentioned things like Bard chat, TBT for the audience earlier, like those are free tools by the way, all of you, if you want to test any of this stuff out, we're talking about some of the predictions, things like that. You can have fun with it. Hey, I need a Taylor Swift song about someone working in hr. It'll do that. I've done it. I just to prove it. But you can also ask other kinds of questions. Zach, any suggestions or tips for them? Anyone out there that hasn't used that yet, hasn't tested one of those out that could help them be more productive, be more efficient, you know, get better use outta their time if they're testing out and using some of those tools, kind of tools alongside them day to day? Yeah. Um, so I'd start with a couple of don'ts, uh, which are, I like, I think one, don't assume that you need it and don't assume that you don't need it, right? So, uh, my 2-year-old has decided that he likes helping me make coffee in the morning. And this morning he was helping me and he grabbed a tea steeper and he wanted us to use the tea steeper, which was sweet. And like we figured out how to make it work, but that was not a helpful tool for making coffee, right? Uh, so like, you know, these tools can be helpful, but we don't just need to insert them into the situation. It's not just that we need to use AI or we need to use these things. We need to understand how we want to use them, and why would we be using those versus something else. Uh, but we also can't assume that we don't. So I mean, it's, it is not, uh, uncommon for me to talk to partners in organizations and realize that somebody's doing some kind of manual data cleansing process that's like, uh, you know, we can take your two hour thing and we can get it down to a two minute automation, uh, pretty easily, you know, just by looking at a, a power query guide from Microsoft Learn, right? This, this not that complicated. So don't assume that you need it and don't assume that you don't need it, um, but, you know, understand what it is that you want to be doing. So, all right, you've got that outta the way. Uh, another big, don't, like, don't pass any privilege data into these tools. Um, there was a, a particular story about a company, I think it was an automaker where someone had, uh, had, uh, one of the services summarized meeting notes for them, and that ended up getting spit out to somebody else who wrote a prompt. And so all of a sudden there was all this privileged information about, you know, plans about, you know, cars that they were releasing or, you know, detailed finances that got exposed to somebody else. Uh, so that's a very real possibility. Um, you know, when you're dealing with things like chat, GBT and Bard, unless you know how to get your own instance of an LLM on a machine and you know, it's not connecting to the internet, just don't do it. Uh, you can still do things like use masking, use mask data or parameterized, you know, uh, you know, if you've got A-U-R-L-A private URL or something like that. Uh, but just be very careful about not passing in, things like that. Okay. So getting most of that out. Well, I guess one other big don't like, uh, most of the things that you'll hear out there, you'll hear, you'll like, see these headlines about, oh, uh, somebody got a better answer out of chat GPT by uh, being kind to it. And it's like they were asking it to play a game and I said, oh, take a breath. It's okay if you make mistakes and it performed better. Right? Okay, well if they update the model tomorrow or it's playing a different game, that may not be true. So the prompt or whatever the guide was that worked for somebody else may not work for you. That could have been very particular to that circumstance. And even if it was general, the model gets updated, everything gets blown up. So all that aside, think about what the points of friction are and where an intern would be helpful for you. Right? So like, think of chat GBT or Bard as an intern and have them do that kind of work and check in with them in that way. And I think that's a great recipe for success. Um, and, you know, starting from there, and you can even go on a little bit, there was a email that I was trying to write the other day and, uh, I, so I pasted it in, I was like, Hey, uh, make this sound nicer 'cause I know that I'm angry, I need to get this sent off quickly, but I'm angry and I don't want it to come from a place of anger so you can make, can you make this sound nicer? And it did a great job of that. That was fantastic, right? Um, there was a talk, another one, there was a talk that I was writing and I, I said, here's the subject area, here's some things I wanna cover. Can you write an outline? I didn't use any of that. I completely redid the outline, but for me, having something to react to and then saying, no, I want it to look completely different, but this is how I want it to look is really helpful. So if I understand what it is I want and I approach it with a thought of like, okay, this is an intern, they're gonna screw up, but it'll still be helpful. Uh, I think that's a great way to approach that. It'll screw up, but it'll still be helpful. I, that may be my headline for the day because that was probably my, uh, my, how people saw me in my job as the HR intern all those years ago. So you'll probably screw up, but it'll, it'll work out. It's not gonna limit a damage that you, cause I, I talk about using it as a, as just a fire starter. And that's kinda what you're sharing here. Like, just use it to, to get something going, whether you use it or not, edit it, change it. That's safer than uploading a bunch of your stuff and asking it to change that or modify that or consume that. So I'm totally with you. Thank you for like, you went like super deep on the, the warnings for people too, by the way. Don't just take it as gospel. Everything else, there's some, some really good recommendations there. I love and appreciate that. So all of you should be testing this out if you're not, give it a shot. We have a really great question from Brenna I'm gonna get to in a minute that actually touches on what you just said before you even knew it was in there, Zach. So I'm gonna, I'm not direct that one back to you just to start off. 'cause it was a, it's a good question. I don't know the answer off the top of my head. Alright. Ah, one last question for you before we get to some of the audience q and a. I've seen some good questions already dropped in here, so if anyone else has one, please throw it in here. I'll take any easy ones. I'll pass the rest of the hard ones to the panelists. They'll be totally okay with that. So Amit, um, we heard some good things here. You talked about outcomes a few minutes ago. We heard Zach talk about some of the risks, like don't do some of these things. What are some of the other things that you think we should be thinking about? Whether it's a risk, it's something around governance as these HR leaders who are thinking about using these tools or even giving, um, instruction or guidance to their other parts of the organization using these tools. Any other risks or anything else in that equation that you think they should keep top of mind as they're making those decisions? So Ben, uh, that's, that's an excellent question and I quite like the way Zach articulated it. Assumption. Look at the assumption and the need. I think that is clearly the driver. A lot of organizations, business leaders tend to think that we need to jump the bandwagon and really look at AI before really trying to understand what the problem they have, right? Mm-Hmm. I think there are a number of broad considerations before evaluating an AI led technology. Some of them, again, broadly speaking, is what is the problem that you're looking to solve? Uh, is it, is it broken? Does it need a fix? Or is it something that you can do away by not looking at technology? Right? That is, that is the most fundamental question, which is often overlooked in, in conversations at an executive level, at a leadership level. So that is fundamental. Secondly, I think the problem itself, do we have a clearly ar articulated problem statement? And why that is important is it clearly then defines the outcomes that are expected, right? If you do not have a problem, if you do not have outcomes that you are expecting, then you are actually going to go, uh, in different directions and that doesn't help. Now, the third thing is we need to understand that AI led technology on models work on data, sanctity, clean data, large volumes of data. If the data is not clean and is not in an appropriate or structured format, then we know what the result is going to be. For example, if you are a healthcare tech company, uh, like Zach was talking about, be considerate about the kind of risk you are gonna get into. Clearly patient data, medical conditions is something that you're not gonna touch upon. And again, uh, if you can look at certain kind of patient profiles, well, the fundamental question that anyone can ask is, is there a valid lawful basis of touching that data? And for what purpose the data can be utilized? And if the answers are no, then there is no question of getting into that conversation. So in a nutshell, it is a good idea to look at the governance protocols surrounding the data. What sort of data can be touched? What is the purpose of data that is going to be utilized for? And last but not the least is, is there a real need for the organization or the leaders to look at AI technology before all the options are exhausted? That is something that I think needs to be simplified, uh, in a, in a very busy world, we often get over engineer and get complicated, but I think simplicity is the way to go. One of my good friends is the head of talent for a large organization. He said our, our chief data officer came to me and said, we have to get some AI was the actual quote that they gave. They said like, I saw an article in the news about our competitor use. They're using this AI hiring platform, we have to get some ai. And his question back to them was exactly what you said on it. He said, what are we trying to solve? What is the outcome we're hoping for? What is the, you know, solution that would make you happy? Oh, okay, well if that's a solution, we may need an AI tool for that. Or we may need half a headcount on a recruiting team, or we may need, you know, someone who's gonna plug these two data sources together and we've got that data already at hand. We should connect those two sources and we're good to go. But it's not just blindly pursuing this path of doing that, no matter the cost, no matter the, the impact, no matter, it's really being careful about that at the beginning. Wonderful. Okay, so I'm trying to parse these questions out that we have already here, uh, in, in line, we have about 10 minutes left on this panel before we transition to our next topic for the day. So I'm trying to make sure we, we use the most of that time to get some good stuff here. So I'm gonna throw these questions out. I'm gonna tag one of you panelists to start each of them trying to align them best I can. But the other panelists welcome to jump in, welcome to share some thoughts on that as well as we go through this. So first question, I'm gonna tag you, Amit, as my first person here. Cecilia said, do you see human resources being more impacted or less impacted than other career fields by ai? What are your thoughts on that? So that's a timely question, and the reason why I say that is, uh, if you look at HR function, the most popular use case is talent acquisition, right? Mm-Hmm. Uh, and then there are use cases like talent development. I think before, uh, talking about impact, it is important to understand, does HR understand how to leverage that technology to get the outcomes desired? I don't think the skill level, uh, there, there's a need for us to revisit the skill required to draw insights and make those recommendations to the business, right? But clearly AI is going to be there. It has just started. I think it's on us to adopt it. Uh, but use, based on what the use case is, not go blind at them. Mm-Hmm. Excellent. Zach, Brad, anything you wanna add to that? I, I would agree. I think there's, there are a lot of, um, tasks that are, are managed within HR that lend themselves well to, uh, to automation. And I think there's that opportunity. I also see it not as being a threat to hr, but really an opportunity to free HR practitioners up to engage in the more meaningful, uh, interesting and, and even rewarding work. So I think, I think that's the right way to approach this, to say, okay, what are the things that we're doing that, that take time and attention to detail that can be automated? Well? And then if that frees up our time, you know, how can we, uh, add value to the business in, in other means that would, you know, support our, our broader purpose. Excellent. The only thing I'll add, I think to that great answers, the only thing I, thing I'll add is do any of you remember back when there was this big debate, do we need a social media policy for our company? I mean, it was probably like 15 years ago, 10 years ago with lots of companies were debating on the internet's a big thing. Social media is a new thing. Do we need a policy around this? How do we manage that? The same thing is true of how and if we're using some of these AI tools. Zach shared a great example of a company, someone did that, whether with with leadership knowledge or without. And suddenly all the proprietary, you know, stuff was aired out into the world. And there are other stories like that where people have done used it to, to fix code and then suddenly their proprietary code is out there. So we have, we as HR leaders have that other responsibility of, we're thinking about the governance or we're thinking about protection of information just like we have to do. If we're locking down other, other things, it's probably a good conversation to have with your IT team about what's, what's possible, what we should be focusing on, what we should be aware of. Because at the end of the day, it is a people thing. And if we're automating parts of people work, we're gonna be in that conversation from the beginning to the very end and looking at what work is best suited for it or not. Not if you're not an expert on it, that's okay, but now's our encouragement to you to go out there and learn more about it. And again, the panelists here are giving you some good tips and pointers on that. Okay. Zach, I wanna come to you for this one. 'cause you mentioned ai, use it like you would use an intern. Like throw things out there. If it gives you a great answer, wonderful. Run with it. If not so much, it's okay. It's just a starting point. Brenna asked a great question about this. She said, AI's great for assisting with administrative work, but how do we balance that with what we might typically give an entry level HR admin? So if we're trying to get someone on our team to build experiences, if we're just using AI for those things, now what happens to that person? So what are your thoughts on that? Yeah, um, so I, some of it I think goes back to the, the purpose, right? And so if we just wanna save time, maybe we do go the chat GB two route or the AI route, right? Because if that's the primary thing, uh, if we're trying to develop somebody, maybe we decide to give them that task because that's good for them. But then we can also think about that more broadly. Is drafting that email or, you know, drafting the policy or seeing what work has been done on similar policies or what definitions have been used, is that what's good for them? Or would it be better to have them go shadow the members of five different teams, uh, and spend two weeks with them in depth, not contributing any value, but learning really, really deeply about how a bunch of different areas of HR work so that they can be a more active part partner on higher, higher functioning activities? That was a really good answer. There's a, I had no idea. There's an organization, I can't remember the name of it right now. It's like maybe Danaher. One of the things they do is when you start there for some of their key roles, like you are not allowed to work for the first 30 days of the job for that reason, we're gonna plug you in here for a couple days. We're gonna plug you in here for a couple days and you're gonna start soaking in the language how we talk about the problems, how we talk about our customers, really getting you steeped in that to your, uh, tee steeper earlier. You steeped in that so that when you, we do set you loose, you have fewer questions and you can hit the ground running than someone who just like you were in another job yesterday, you're in a new job today. Trying to keep all those things together can be really difficult. So I really think that's a really good example that, that could make that work more meaningful, more relevant, more helpful for that intern, that entry level HR person than us asking them to go and type up another termination letter or something else. All right. Wonderful. I would just, I would just add to that, I think there is a different way to look at it for that question is the kind of skill, right? I think in the next 10 years, 15 years, the jobs are completely looked very differently, right? As a result, a lot of organizations are focused on building skills in the individuals, right? Again, Zach was talking about the same thing. Uh, well, we need to look at what kind of skill this person can be developed and where their career can be be headed towards. I think skill building would be an important dimension to look at, uh, rather than sort of looking at purely from a work perspective or a role perspective. Wonderful. Awesome. Thank you for chiming in there. Okay. We have just a few minutes left on our last question here and then we'll I'll wrap us up and transition us off to our next section here. But Brad, I have not left you out. I've saved a good question that I think you're gonna, you're gonna love. So Deirdre said, how can AI support performance management or performance evaluation processes? Ideally, all of us probably have this goal. Deirdre, ideally I wanna make it better for the manager and the employee. Yeah. So any thoughts on how AI fits into that? Yeah, so, um, you know, it's interesting when we analyze survey data and performance data, you know, we, we often employ, um, quantitative analysis. We're looking at correlations associations, but the reality is that data can be highly subjective. Um, and so I think we need to one, come at it more with a lens of, of qualitative research, kind of understanding the lived experience and understanding the unique situation and needs and context for both the manager and the individual. One of the opportunities that I see coming down the pike is really integrating more data around, um, you know, strengths, um, where that person is in their career, uh, their career trajectory. I really like, um, I believe it's a radical candor Kim Scott talks about, rather than looking at performance and potential, she looks at performance and growth trajectory. I think that's something that can help set managers up for success in conversations around performance management is to understand, okay, do we have high performing low growth trajectory people that Scott describes as, um, the, the rock stars, the rock steady individuals, um, and what are some of the prompts that we can give the manager in that performance conversation? And how is that different than a high performer, high growth trajectory individual that's really looking for growth and, and more opportunity kind of beyond. Um, I also have seen opportunity, or I've seen technology where, um, managers are able to engage with a chat bot to practice some of those difficult conversations. And, and they may be a good conversation, but they can be awkward and managers may not feel equipped for it. Um, and the example that was given is they said, if you're looking at special forces, if you're looking at athletes, if you're looking at anybody that performs at a high level, nobody walks out onto the field or out into an environment just to perform right out of the bat. Everybody is going to have some, some practice before that. And I think AI is gonna give managers that opportunity to practice, engage, get feedback on their own communication, and then go into those conversations more confident, more comfortable, and ultimately more effective that's gonna benefit both them and the people that they're leading within the organization. So, um, you know, it, there's a measurement component. There's an analysis component. I also think there's going to be an equipping and enabling component that's gonna be powered by AI in the very near future. Great. Great answer. You talked about like preparing the manager piece, especially, I'm thinking about, I ran across the chat bot a few years ago that was meant to help employees when they had, um, they had a problem with like discrimination or harassment, so the negative side of this, but meant a problem many times. They weren't very good, like the bystander thing. We're not good at seeing what's happening around us and capturing all the elements of what exactly happened, what happened next, who else was there, who was a witness, how did I feel? What time did it happen? Like, all those kinds of things. And the chat bot was built to structurally take someone through that experience. So they captured all the details, whether they decided to report it to HR or not, but it helped them to capture that. And the same thing, I think helps that manager think through it. It helps the employee think through, oh, I forgot. I actually did, you know, had this big change this year in my job duties and I knocked it outta the park. Yeah. And because of that, I wanna advocate for, you know, 5% or whatever else in the, in the, you know, when it comes to performance and cough and everything else. So. Wonderful. Okay. I know we are here at time, Zach, I'm it Brad, each of you, incredible, incredible, uh, minds in your own right. Thank you for joining to make this an incredible panel. I appreciate you for sharing, for encouraging the audience, for giving us some good practical ideas and takeaways. This has been such a joy to start this day off today with ai not in place of humans, but to support the humans and make us all more effective HR leaders.

‍