Rethinking Work for the AI Age with Ketaki Sodhi

Speaker A: Welcome to the Edge of Work podcast. I'm your host, Al D. This is a podcast for leaders who want to make sense of workplace trends and are looking for new ideas about how to lead people and grow their business in a changing world of work. During each episode, I'll bring you the latest experts, researchers, founders and leaders to share new and unique ideas, as well as actionable advice around attracting and retaining talent, developing people, and building healthy and sustainable organizations. Welcome to the Edge of Work. Today's guest is Keiki Sodhi, who is the head of AI enablement at Moody's. Keiki, thank you so much for coming on the Edge of Work podcast. I guess before we start the conversation, I would love to have listeners learn a little bit more about you and your role. So would you mind sharing a little bit more about what your role is and how would you describe the work that you do?

Speaker B: Yeah. Firstly, thank you so much for having me. I'm so excited to have this conversation. So, like you said, I lead AI, ah, Workforce enablement at Luliz. And the short description of what I'm responsible for is essentially making sure that as AI becomes more and more ubiquitous in the workforce, in the work that we do, we as an organization are able to absorb that intelligence and get value from it. If you think about the intersection at which I sit is really the intersection of talent, technology and transformation. How do we use the technology that we have at hand to amplify what our talent can do and really transform the way in which we internally generate value as a business?

Speaker A: So one of the reasons why I wanted to chat with you, in addition to just enjoying some of the conversations that we have, is that I know that you follow the AI conversation as closely as you can, just given how robust and how much is happening. I guess we're recording this in the summer of 2026. We'd love to just get your general thoughts or observations on the state of AI right now in the workplace. I know broad question, but when you think about that, what comes to mind or what are you thinking about?

Speaker B: Percy, I love that you're saying we're in the summer of 2026. We are specifically in the middle of July because this stuff changes so quickly when I think about where I sit in my background and really what's top of mind. There's so much going on in the AI space from the technology perspective. But if you think about my background, I am an organizational psychologist by training. I've spent my career in talent management, people analytics, and a lot of the open Questions right now that organizations including us are working through are beyond the technology. Like the technology, we have a lot of good technologists that are looking on getting that to work. But there's, it's the behavior, the incentives, it's how do you rethink workflows, how do you rethink about people's identities, culture, measurement? Those are all people science problems. And when I think about what's top of mind for me as I follow this space is the gap between what I like to call the capability absorption gap, where the speed at uh, which model capabilities is moving so fast, it's compounding so much faster than most organizations have the capacity to absorb and do something with it. It's a gap between what's possible technically and what's actually deployed and how work gets redone, I think continues to get wider. And the work that a lot of AI enablement transformation leaders have to do is figure out how to make that gap shorter. And if I were to be a little bit more specific in terms of the types of things that I'm trying to stay on top of or things that I'm observing, especially when I think about large scale enterprise AI, I think the first is at the beginning of this there was a lot of focus on individual AI, right? How do each of us upskill and learn and become better users of this technology? I think now we're moving into more like organizational AI. I have a, uh, really strong person on my team that calls it institutional AI. How do we think about moving from each of us is getting value for our own work to a team can now orchestrate work using AI. I think the second sort of big theme that I'm seeing is this move from what we call a thousand wildflowers to organized farming where there have been so many experiments that have been blooming across the enterprise over the last year and a half across the board. When I talk to my peers, that's the theme. But the hard work I think in this 2027 or 2026 is really thinking about how do you take that grassroots energy, that bottom um, up to your mixture and turn it in tune highest value use cases that work, making sure your data quality is there, making sure the bottleneck as it moves from his creation to actually like verification and judgment. There's enough capacity to absorb that move the bottleneck and then also rewarding people and incentivizing them for doing the right things as we move along this journey. And I think that really is a systems thinking problem that brings together people, teams, technology, teams Finance teams, communication teams. I think that's really the crux of I think what a lot of enterprise young readers are thinking about right now

Speaker A: that resonates at least with me. And to the first point you made, particularly around that individual productivity piece or thinking about individual activities or actions and trying to get it beyond that, I think that what, what strikes me about that piece of that conversation, I think some of it really starts with just also this reminder that work is a collective effort. Right? And for the most part it is hard for one single person to be able to actually do something solely on their own, particularly in the large, large mid size to large enterprise. And on one hand it is hard for anyone to get value out of AI if they don't understand how to use it or if they refuse to use it. And so there's a certain amount for sure of uh, this is what it is, this is what it looks like, this is how you can use it. And this again, this reminder that a lot of the real value I think really does come from those workflows that are integrated or are cross functional or do move from one person to the next. Because what good is it if it's great that you can produce 50 PowerPoint decks now when you can only produce 10 before? But if what's going to happen to those other 40 is that just going to get the bottlenecks? All the bottom. If you don't get beyond those individual things, I think the realization is, oh, the bottleneck just shifts on to someone else and it doesn't actually really yield to any other kind of productivity. It sounds like you did a kind of resonance with you.

Speaker B: Yeah, that's, that's exactly it. And I think the framing of, I think individual readiness, skilling, all of that is almost a floor that, that's the foundation on which everything is built. But I think the system that you're in sets your ceiling and so you need the floor to build on, but how high you can go and how deep you can build, I think that's where that people readiness of course, but you need leader readiness, you need technology, you need outcome readiness and all of those sort of play the role in the system to set where you see leaders essentially.

Speaker A: I like that framing and I think to your second point too, again, it's a good thing if you have people experimenting, trying things in the spirit of wanting to see what works. It's a good thing from a perspective of a lot of numbers of not having just one person in your organization trying to experiment with AI and having lots of people to do it and to at some point to get some kind of scale. It does need to move beyond and focus is a good thing. But the, I think the always the, the challenge always is. And I think you articulated a little bit with that, thinking about this in a systems kind of way of how do you thread the needle on that of how do you thread the needle or find that happy medium of yes, we very much do and want to a certain degree everyone being involved in this. And it's not just one team or it's not just one function. And we've got to think about this in a thoughtful way, in a systemic way, because we can't just have everyone, which team doing all these kinds of things because uh, that's also not helpful to trying to get some of those systemic benefits as well. And I think that's, that's the struggle that I've seen too.

Speaker B: Yeah, 100%. And I mean before we got uh, on we were talking about that, the World Cup, I think soccer is an example of some of this. Right. It's an aggressively team sport. You're working in that system, but you meet people in different roles doing different types of things, things to make the outcome actually land. And I think that sort of like different roles be played by different people at different levels. And having the readiness to do that like that to me feels like a much closer analogy to a lot of the more like individual focused work that I heard of.

Speaker A: More, I'm curious, as someone who follows the broader discourse about the AI in the workplace conversation as well as someone who stays pretty plugged in, I'm wondering, as you've read and talked to your peers, et cetera, are there things that are not being talked about that you think deserve more attention or you deserve more articles? I would go to say is deserve more podcast episodes about that, but the last thing we need is more and more people with more and more podcasts. But you get the general nature of the question I'm asking. I'm just curious, as someone who is plugged in, I'm just curious to know from your perspective of what isn't being talked about that maybe deserves a little bit more thought and intention.

Speaker B: Yeah, I uh, really like that question because I'm really plugged into a lot of what's happening across the different areas of the financials, of all of this, the token economics, the technology. But I think the most under talked about part of all of this, and it's always a, uh, it's always a footnote in things like you need your data to be in order. But I think what that means is so different in today's environment where I think that like organizational content layer is in a lot of ways the crux of where the gaps between, hey, I have a cool agent that works in demo versus something that actually works in production lives. And I think that most organizations knowledge is tribal. And when you think about data, I've uh, actually had people say to me like, oh, we actually don't have that much data and they're thinking about numbers on a spreadsheet. But when AI can crunch meaning and context and bring that all together, everything is data. Right. And so that organizational knowledge that is usually tribal lives in people's heads. Slack threads. Humans just know how to navigate it instinctively. But I think that how do you organize and then harness that across different fields of an organization? I think that's a really key part, uh, analogy that I really like around this is you might be familiar with that classic exercise of have you seen the videos where they ask children to give like parents or teachers instruction on how to make a peanut butter sandwich?

Speaker A: Yes, yes.

Speaker B: The children will say, put the peanut butter on the bread and the parent will just take the bottle of peanut butter and put it on the bread.

Speaker A: Yep.

Speaker B: That is exactly how it feels right now at stale when you think about the same data in context being used for humans and agents, agents sitting on top of your messy enterprise data in the parent that's putting the jar on top of the piece of bread. We need that layer to be so much more organized in a way that makes sense to a machine, uh, as opposed to a human. A machine doesn't know that my spreadsheet that says execdex final v3 17th of July is the actual final one versus finally two. The machine doesn't know what a client means in the system or why March numbers might look weird. All of that on top of your actual raw data, I think is where the power lives and it's the ground within which your agents can grow. And I just don't think we're talking about it enough.

Speaker A: I like that one. I want to share two anecdotes. I'm curious to see if any of these lands. So the first anecdote I wanted to share, I was catching up recently with a friend of mine who works in, uh, talent. It's a talent partner role, but specifically it's within the IT and engineering organization of the company that they work at. And I was just asking them, I was like, hey, what are you doing tomorrow? Well John, your priority was for the week. And one of the things that they were saying was that they actually had on the calendar for that day. It was almost like a company wide day or for at least for the org where their entire job was, it was a documentation day. And what they were literally doing was what you were mentioning in terms of documenting workflows, documenting processes, cleaning things up. But part of that was because they are at least in the IT and engineering parts of the organization, maybe less so in other places using so many agentic workflows, using, using all these sorts of things that the value of what you get out of that is only there when you have that context. Right. And also in that matter, the real value in what some of the technology can do actually comes from that tacit knowledge that sometimes lives people's heads or around the work, but isn't actually showing up or cannot be ascertained by just looking or wheeze at a deliverable. So that's the first thing they wanted to say. And then the second thing I wanted to say and offer this up. Someone was asking me just almost as jokingly around, how do you actually tell if a leader actually, or uh, an organization for that matter who is posting on LinkedIn about what they're doing in AI has actually done anything meaningful as it related to AI? And the only thing I could think of was like if they actually post about all the work they're doing around their data cleanup. Because what you can do with this is only as good as the data that you have. And so it's like the, almost like the heuristic intel is like of if someone's actually done anything is how unsexy of the work they've actually done.

Speaker B: Yeah.

Speaker A: So I'm wondering if you do either of those two ants.

Speaker B: Yeah. Ah, no. Both of those resonate a lot. And I think that IT example is actually a really good one because in some ways this is a chicken and an egg problem where a lot of times people will say oh, why software development so much ahead of other use cases when it comes to we have our software engineers, our product teams, they are really leaning into this inner raid is very impressive and I really think it's chicken and an egg in a couple of ways. One being yes, coding agents are the frontiers like capability agents harnesses that we have which only make it right for this. But if you think about the engineering culture, it is a culture of gasification. If you think about the way in which engineers get worth a dining, it's in GitHub it's in papers you always have somebody reviewing before something goes into or gets merged into code. And that sort of muscle. Yeah. Is what a lot of other teams that are non technical. Ms. Yes, I was trying to talk to one of my non technical leaders about this and it was a complete mishap to find a better way to say this, but I said something along the lines of I would imagine a world in which in the future it's not just our tech teams, all of our teams live in a GitHub like environment with our COEs become older MD files. And that's not a exciting new direction when you say it like that. But I think that muscle of maintaining documentation, consistently updating documentation and having somebody actually be responsible for it. On my team we're doing an experiment right now who are building out arch infos and we've had this concept of a librarian now where different people on the team take turns on a weekly basis to prune our sort of like contact layer or our set of data and information as the human librarian on top of everything that our agents are curating for us in our team os. But ah, that's like a muscle that I think a lot of non engineering teams have to build and it's not easy to build that when you haven't had to operate like that for a really long time.

Speaker A: I really like that example and maybe just to make this real for folks out there who are listening to this and are like why would someone bring a librarian into their team? Could you talk about at a high level at least just this concept and idea of why something like a librarian, particularly for folks out there who don't work in an engineering or IT or product function, why something like a librarian or the things that a librarian could do could be really helpful and valuable. If your team is starting to think about doing ejectic workflows or even not even ejectic workflows is just starting to try to go beyond just individual productivity with AI tools. What would be the benefits of it or why would it be so critical and important?

Speaker B: Yeah, I think the framing that's really helpful here is the cold start problem that we're trying to solve is imagine you have an intern that it's their first day on the job, you want to give them a real problem to solve. All they have access to is a computer and all your SharePoint files and no human to ask, no way to know outside of let's say your Internet what is and you've given them a pretty exciseable task to do, could they do it with all of the information that's within your systems? And the answer almost always is no. But when you think about agents, that's exactly what they are dropping. This fuzzy machine that can non deterministically process information and do things with it. And in that environment there's two things that's happening. So if I think about the architecture of our team OS as an example, but it sort of mirrors what a lot of people are building right now. It is. And Andre Karpi is one of the really big thinkers in AI and he has this LLM wiki that we borrowed from and then there are a few other frameworks that we borrowed from, from Stitch Together what we are experimenting with. What you can imagine is it's a three layered system where the bottom layer is what we call our capture layer. This is where anything and everything that's happening in a week, in a day, you throw it all in. We have a, uh, manual part of that where we have a teams channel where nobody monitors it, but we just throw in information if it comes our way that we want everybody on the team to know. There's also a, yeah, declare on it where on a daily basis, based on emails and in chats and stuff, it synthesizes all that information. All of that goes into a second layer, which is the link consolidation layer. So we have workloads that take all of that raw data, uh, and organize it into the different projects that we have, the different work streams. Things get updated. And so that's sort of like layer in which you grab things from when you're trying to figure things out. But the layer that sits on top of that is agents, our librarians and us. And the librarian becomes really important because you can imagine, like, imagine going to the New York Public Library. You have m hundreds of thousands of books. How do you navigate that to know what's the most important thing to pull? And also how do you cull things that maybe is not that important? And that's the role that our human librarians play. It might not be something that's required over time. Right now we need that QA layer because we're still testing and experimenting with this loop. But the idea is that you're pruning from this massive messy collection of information and you're quality checking to make sure that the agent that's synthesizing and capturing your information is doing it accurately in the context of the work that you're doing. And then over time it might go from weekly to monthly because you have enough confidence, you know, you start getting good enough. But then that's that layer that we can query into. And so if I want an update on the thing my team is doing, I can just go to autocopilot and say, hey, did this person have that meeting and what came out of that meeting or was a decision made? I don't have to bug somebody for that. I don't have to remember that. I can go and query our sort of like knowledge base and context on that.

Speaker A: Hi everyone, Ald here. Thanks so much for listening to the Edge of Work podcast. I hope you're enjoying today's conversation. In addition to hosting the show, I spend my time advising, coaching and partnering with leaders and organizations who are navigating a rapidly changing workplace. If you or your organization are focused on helping your employees and leaders lead through change, strengthening their human skills in an increasingly technology driven world, build greater adaptability or navigate the AI change management challenges, I'd love to connect. Whether you're looking for a keynote speaker for an upcoming leadership event, a leadership program, or just want more hands on support through workshops or coaching, I'd be excited to learn about you and your goals and explore how I can help. You can find my contact information in the show notes. I'd love to hear from you. Now, let's get back to the show. I think that's fascinating and I think it's a, uh, in my view at least it is a window into where this conversation is headed, particularly around how do you actually reinvent the way that you work using AI tools. But I want to bubble up for a second or pull up for a second just because I'm sure there, uh, might be some people listening to this thinking to themselves. We're definitely not doing that in our organization. And so I guess maybe just to ground this in set context, if you think back to maybe where we started around AI enablement or workforce, really thinking about AI adoption and things like that, and let's reset the clock on, let's just say for the sake of this conversation, like two years ago, uh, versus to where we are today. I think what you're describing is, and I know you mentioned you're experimenting with this, so it's not like you've necessarily figured this out, but when you go and talk to other leaders in your, whether it's in your industry or just your peer set who are also thinking about AI adoption, AI enablement, would you say that what you're describing is something that is widely held? Are we still in that area of still Trying to get people to start using these tools, start getting permission from a compliance or legal or IT perspective, actually getting people licenses. What's your general just pulse check or sense of uh, where are we collectively? I know that's a broad framing, but where are we collectively in this conversation around AI adoption and enablement?

Speaker B: That's a really good question. I will break that down into a few different buckets that I think is helpful. I think the first is I think the companies and teams that are furthest along this journey are the ones that are either startups or frontier model lab companies. If I talk to my peers at ah, anthropic OpenAI, they are inherently working in these ways that are AI. Ah, first I know the companies like, think about Microsoft, Google, they've been on this journey for several years now. They are starting to make this a part of their day to day. I think the second category of companies that I see are companies that have more of a risk of not doing it than they have a risk doing it. And given how many startups are popping up in financial services healthcare, I think there's a pretty strong peer set of financial services companies that are thinking about this very holistically. And so thinking about the tooling access was a problem of 2020 factory tracing. Now it's really about how do we optimize, how do we make sure we're getting value, how do we build the system and the ecosystem around it. That said, I think there's still a lot of companies that are stuck on that first up dealing side of it. Let's give people tools, let's get them to learn. And if I were to just think about my peer set, I'd say it's probably like a 30, 70, 20% to 30% of us are thinking about the broader system. And that comes from, it comes from really visionary, I think leadership. And so I think the thing that's common across all of the companies that I talk to that have this more holistic sort of systems lens are the ones where they have their C suite leaders, their boards really saying you should be leaning into this. And we are 100% wholeheartedly behind this. We have to do it responsibly, we have to do it safely, but we have to do it. And then I think the companies that are a little bit stuck in that first stage of this are the ones where there's a lot of energy, there's a lot of interest, you're sort of seeing your peers do it, but there's not enough directed guidance from leadership to really go all into it. That to me is the number one sort of signaling difference between them outside of the operational. It takes a lot to actually get this done. But that leadership buy in is a huge unlock, I think, for a lot of my peers that are further ahead, for sure.

Speaker A: Again, like even to your point about the frontier firms as well as the AI native firms, when you think about leadership in those cases, right. It's. If you work for an AI native firm, the leadership conversation is just going to go a little bit different because.

Speaker B: Different conversation.

Speaker A: And so. But that in turn just makes it at least with the confines of should we be using these tools, should we be thinking about working in a different way? It's just a much different conversation, if any at all. Then you, if you're working at an established firm that has been around for 25, 50, 100 years and has a lot of the trappings that come from it. But I am curious though, uh, on that notion of, I guess the best way to frame this, particularly for more established firms, is this idea of what is that role of leadership in that respect? And I think what's hard about this is that I do think clearly leadership plays a role in this, whether that are executives. Oh, whether that are everyday people leaders. But I'm wondering if you've thought about this or this has come up at all around. How do you try to set and define or create clarity around a future in this period of uncertainty in the sense that, let's just say for a sec that we agree that we can't keep doing what we're doing today, but also, for that matter, it's hard to pin towards a future that really is still forming as we speak. And so how do you maybe thread that needle in terms of being prescriptive enough and again, acting with confidence enough that people actually believe what you're saying, but also to the degree that, as we talked about in the beginning, part of why we're timestamping this conversation is because so much is changing so fast. And how do you not get out over your skis, if you will, of either going too far into the future or maybe being too short and then having things kind of get better or shift under your feet?

Speaker B: I think at the outset of this, I'll say I think it's impossible to perfectly thread that needle at this point. This is a space where I think the ground is shifting under us on a daily basis in a way that when I talk to my peers in the way that I think people not experience for a very long Time, right. And we're in the early, early. A lot of people say we're in like the 1999 of the dot com era, where it's like really early as the technology evolves. And even as I talk about things like the context layer, we're in the hobbyist phase of that. And so to your question, I think there are a few answers I think are really simple and then there are a few things that are a little bit harder on the simpler side. And when I say simple, not easy, but they're simple. One is culture. We've been talking a lot about the geek way and how we lean into this culture of continuous iteration and continuous improvement and strong candid feedback to really build a culture where we can, we're resilient to change because I think it's not realistic to think things are not going to change. So how do we build a company where leadership down, we're resilient to that sort of change and have the culture to be able to absorb that. I think the second, more operationally that I think has been really valuable is as things have changed, having a strong operational rhythm where multiple options can come together and have this full conversation is really critical. What I mean by that is I think a lot of companies, I see some of my peers that are in chief AI officer type roles where AI is one person's responsibility, is your outcome versus a few other types of teams. And I think Moody's a great example of this, where the responsibility for AI is distributed amongst a lot of people in leadership in different functions. But there's an operating rhythm that brings those leaders together to make important decisions. And there's enough work done to set up those operating rhythms so that when things change, you can quickly come together, have that conversation and go back out to really implement the decisions. Think about very recently, this was March, April, Anthropic changed their pricing model for enterprise, where it went from seed based to usage based. That's a huge change to make in the middle of a fiscal year, something like that. That is not a single leader decision. That is a finance, that is a people decision. That is a technology decision. There's a lot of different things. It's a communications decision, it's a change management decision. And so having the operating structure to come together and do that quickly I think is really important. And then the third thing I'll say is having enough of a, uh, relationship with the business makes a world of a difference. And so something that we've been really investing in is the hub and spoke model for AI where yes, we have enterprise enablement by partnership across our uh, corporate functions that set the ceiling for what the company can do. But the real work is happening on the ground by the business, by the subject matter experts. And having that sort of line, direct line of contact with every line of business that matters I think makes it such that it's not a top down, it's not a bottom up, it's a true partnership. And so this is, this is agnostic of AI, right. But this is for any change, I think these operating rhythms are really helpful. When I think about AI specifically I think that is where I really like the Jeff Bezos one way doors versus two way doors framing. Um, what are the things that we invest in right now that are, even if the bull case for AI doesn't pan out, will not be a waste data infrastructure. Having better data, having seen our data, even if we're in the bigger case of AI is not going to go to waste. So let's continue to treat those as places where we invest a lot of time and energy to have that decision grade intelligence and then places where we are a little bit more, maybe we don't have as much conviction. Let's keep the door open from like a, uh, if we take the lens of strategic foresight, let's treat that as a more wide lens and have a few that we're preparing for and we can narrow it down as time goes by, but not narrowing that down. Your list of options too early I think is part of that, like management of AI specifically.

Speaker A: I think one of the other things that we had talked about earlier that I did want to get back to a little bit was this idea that you had mentioned. I think someone in your team had talked about institutional AI, uh, in terms of how do you go from just getting people to do individual workflows or individual tasks and activities and move to something that's um, much more systemic than that. And that being like, that's when, if you can cross that chasm where you can start to enter that conversation, that's one of the ways you know you're going in the right direction. I'm wondering how do you know you can get there or what has to be true or what has to happen to be able to elevate that uh, conversation past, hey, it's great that you're using quad and you're using things for yourself. Is it simply just that they're starting to work more collaboratively with cross functional stakeholders on these things or for other people in a particular Given workflow, et cetera or what are some of those other ways that you might know that you're actually starting to move beyond just oh yeah, we have people who are capable of using this and they're logging in and using it and whatnot.

Speaker B: Yeah, that's a good question. There are a few different things that come to light to this. I think the first is outcomes are not tied to an individual. And if you think about a lot of times when you have roles that a lot of people occupy, very often it'll be oh, I have the, I have a great recruiter, so my outcomes are great versus oh, you have a uh, maybe not as experienced recruiter, not a great recruiter. And so your experience as another business function or leader is not as good. I think when you have through workflow redesign in some of these or institutional value that's coming out of it, it starts to even the playing field across similar roles in the same team because then you're using similar types of data, using the similar types of workflows, your check ins look very similar to folks doing similar jobs. So a, that's one knowledge survives a specific person, expertise survives beyond a specific person and it becomes more about the role and the workflow as opposed to an individual person's excellence in a certain area. Not saying that I think there's always going to be rock stars in certain places that are going to shine, but it increases the floor for everyone across the board. I think we start to see that for any second is the shape in which work starts to happen looks different. And I know I sound like a broken record because I'm going to go back to the data but the work starts to look a little bit almost think of it like a horizontal edge where you have a foundation of data that is really strong. On that foundation of data you have a set of workflows or processes that are built and then a host of different people could tap into that and make usage of it. And I think we're a little bit, there's some work to do to get more organizations and work teams to like that. But I think if I think of the shape of teams at uh, organizations in the future that are doing this well, it's going to look like you have a central data and context layer. You have a few deep experts that are not just doing expert work but are uh, building expert systems. And then you have a whole host of generalists that can flexibly tap into those efforts to compose what they need to solve a specific problem or outcome. And I think the last thing I'll say is measurement will move towards outcomes rather than usage. And I see this, I talk to my peers. We're seeing this in areas where when the work is really transforming how work gets done. It's not about, oh, we're using AI. 80% of our team is using AI every day it's your shipping faster, your time to PR is faster, or your customer experience NPS is getting better. And so as things actually start to absorb within workflows, it makes it to a point where you can actually measurably see it in your outcomes and not just your activity

Speaker A: maybe. Last question before we wrap for the day. We were talking a little bit earlier about this idea of, in, um, even related to what we just talked about, of ah, this idea of things that are more systemic. Right. Things that are broad, reaching far across the organization. And we haven't spent as much time talking more about hr. And so I'd love to maybe talk about that for just before we end for the day. I'm curious as you take a look at some of the more systemic talent practices or processes or workflows that we have, I'm curious, particularly knowing about how different they could look as a result of AI, uh, transformation. What are some of your thoughts there, or even if there is one in particular that you're keen on or that you're really curious about, or maybe even that you're starting to think about how you might do this differently? Any thoughts about just in general how some of the things that we've long held as really important and really critical in the work that we do, how could they start to look differently as a result of AI, uh, transformation?

Speaker B: Yeah, I fit in the people team, right? So this is a really big part of our conversation when we think about talent strategy. I think there are a few areas where this starts to become really important. The first I think is compensation incentives, your performance management structure as a whole. If we think about most organizations performance management structure right now, or their generals, or how do you reward people for doing a good job and what it means to do a good job as you go from junior to senior, you start paying people more for leverage and outcomes, right? And so when you have more junior talent, you're often seeing hours worked. How many tasks are you completing versus if you think about some of the more senior leadership execs, it's more about what are the outcomes that you'll be able to deliver with the resources that you have at your disposal. I think a fundamental shift we're going to see is that getting trigon to lower end to the organization. So as more people have agents and AI at their disposal, measuring and paying and rewarding people for using those resources to drive actual outcomes and using it as leverage as opposed to for activity and hours, I think that's going to get deeper into the organization. I think if we think about things getting deeper into the organization, it's also a very basic thing is we're going to have to really rethink how junior level employees build judgment if they're not doing the graph work of every day. How do you create the artificial experiences to give them enough reps to build judgment to be able to make decisions and be your senior level management one day? And then um. The last thing I'll say is I think there's a lot of discourse right now around the role of managers and middle management. I think there's a lot of ink pieces about middle management going away and I think there's a lot more warrants that HR leaders need to account for when we think about the managers. Because if you think historically management served a couple of purposes, right? Management served the purpose of it's an information transfer vehicle from senior leadership all the way down. It sets context for people. The second is it often serves as a hub for judgment and knowledge. You spend 10 years doing the job implicitly there's an assumption that you know what the right thing to do is. As you'd said, strategy for the future. And then the third is this idea of a manager as a coach. So you're playing a role in the development of the future of the talent. I think we're going to move towards a world in which that ladder becomes a lot more important than the role of a manager. And the first few might start to feel way and so that like composition of the role of a manager, becoming more of that coach, creating the teamwork environment, being a coach, helping people grow and develop I think will become a lot more important. And that information flow sort of like let's organize the team and have the context. I think that can very much be handled by more layers of data and AI. And so I really don't think middle management is going to go away in the way that a lot of our think pieces are speaking uh, on right now. But I uh, do think that composition almost like the consulting model of having somebody that's responsible for the project work that you do and then there's somebody that's responsible for you and your development as a person, I think that is going to become a Lot more. A lot more important, especially when we think about junior level talent.

Speaker A: I appreciate that. And that would actually be one of my answers to what is not being talked about, that should be talked about, which is going to the second or third level of that manager conversation. There's a lot, as you mentioned, being talked about of whether or not they're going away or not. But I think there is a real opportunity to get more creative and think more expansively. Even just the frame that you gave about the consulting model, I think is something that needs to be talked about more because I think the other reality of it is like it is still a little mind boggling to me that give or take, that collectively we've decided that we more or less have one kind of org model for how we think about management for all the organizations that exist. And if we just take the principles of just differentiation just at face value, you would think that someone would have stepped outside of the bounds a little bit to try to differentiate a little bit to say, hey, we're actually going to approach this in a different way. And I think that is one of the opportunities here to the point that you made of actually trying to really rethink some of this because it could lead to different types of ways you approach this. Like there might be a cadre of organizations that use this kind of approach where they do break out the different roles of a manager one from the other. There might be some that choose to go the way of what some of the tech companies are doing right now where you just don't have that. And some people might not choose to go to one or to other. But it does feel to me that we've kept things a little shortsighted in terms of not actually really exploring the frontiers of what could be possible. And so that to me feels like an opportunity among many others. K Soti, thank you so much for coming on the edgeworth podcast. It was a pleasure speaking with you.

Speaker B: Thank you so much for having me.

Speaker A: Hi everyone, ALD here. Thank you so much for listening to the Edge of Work podcast. If you like what you heard, I encourage you to share the episode with a friend as well as to head over to Apple Podcasts to leave a review and let us know what you think. I would be forever grateful if you did that. I would also love to hear directly from you about what episodes you're listening to or any suggestions you have for how we can make it better. You can find me on LinkedIn.