How McKinsey Is Rewiring L&D for the AI Age: Heather Stefanski

[MUSIC PLAYING] Hello, and welcome to Workplace Stories. The podcasts where human capital experts share their real world experiences and insights. Hosted by Red Thread Research, Workplace Stories surfaces the latest industry trends, innovative talent strategies, and best practices for success. Tune in for practical advice upon a best transform your workplace. Hey, welcome to Workplace Stories. I'm Danny Johnson. And I'm Stacey Garr. And we just got off the phone with Heather Stofanski, who's the chief learning and development officer at McKinsey. She-- we talked a lot about the future of Ellen D. and some of the things that she's doing inside her organization right now. Stacey, what's stood out to you? Look, first thing that stood out that Heather said that I haven't heard from anybody else yet is that as we're designing all these various agents that are going to do all these things that we used to do, that we need to not just be focused on productivity, but also on development. So this idea that potentially instead of the agent just going and doing something, it could ask us questions that are designed to enhance our knowledge or enhance how we think about the problem, et cetera. So development and productivity together in agents and thus a need to have learning folks actually building agents with the technology folks. So that was one thing. Another thing she talked about that I really liked was this idea of purposeful apprenticeships. And so it was making sure that there are yes specific systemic things that you want to encourage developments or things like rituals or performance check-ins and the like, but then also resources and even agents that can support people as they are developing. She gave this example of an agent that a partners had developed that gave early feedback to people so that they could get that quickly before they even came to the partner. So that was the second thing. And then the third thing was some thoughtfulness around deciding what to let go. We talked about whether baby consultants need to be trained on how to align columns and colors and all the rest of that is that still necessary. Well, you still want that, but an agent can do that. And is there value in the human doing it? Probably not. So when can you stop teaching it? And so that is, I think, one of these questions of what do we let go? What do we keep and when do we make that transition? What about you? Those are excellent points. Three things that sort of stood out to me as well. First of all, I love the way that she thinks systemically like we do. So she's not trying to change each individual person of a Kinsey even though she is trying to change each individual person that I can see, but she's trying to do it through systemic change of how they think about and how they develop people, which I thought was fascinating. She also, we asked her specifically about the skills her L and D team is going to need. And she mentioned the AI skills and the need to reason and some of those. But we also sort of landed that the same skills are necessary. They're just applied in completely different ways. So instructional design, designing a skill that is then embedded into an LLM to help you sort of develop as you're working is a completely different way about applying that instructional design, but the skill isn't different. And so it was interesting to sort of be like, yes, those these skills are sort of what L and D does well. And nobody else in the organization does them. How do we embed them into the rest of the organization? I also love the fact that she's embedding her team in workflow design. So she's taking her people and putting them directly into teams that are developing new workflows and thinking about that, which is the strange for L and D organizations that usually sit on the side of, in their silo and not ever leave. That may be think of one other thing that she mentioned, too, which is how we're going to measure L and D differently. If they're in the workflow and the impact is in the workflow, I'm not going to spoil it. But I think that idea of measurement of the effectiveness of the L and D function is super interesting. Yeah, it definitely helps to change. She also had a really interesting conversation about the tech stack. She and a couple of others are doing some really interesting work on what a learning tech stack looks like. And I will spoil this. Maybe there isn't one. Maybe we're using technology that already exists in the organization. And then finally, I love that when she was talking about developing people, she wasn't talking about just AI fluency. And she actually hardly mentioned AI fluency at all. She was talking about developing these other skills that are incredibly necessary for McKenzie, like problem solving, systems thinking, a medical admission, and judgment. Those types of things as where she needs her organization to focus and how we do that using AI to drive larger changes faster. Who's a really, really interesting conversation as we knew it would be? Please give it a listen. And if you like it, please recommend it to your friends and follow us. Thanks. [MUSIC PLAYING] Welcome, Heather. Thank you, Danny. It's great to be here with all of you. We're excited to have you. We usually start with what we call the first day question, which is what-- just tell us a little bit about yourself to get going. Yeah, absolutely. So hi, everybody. My name is Heather Stephrance. I am the shirt-toe learning and development officer at McKenzie. So what that means is I am both responsible for all of the formal learning and training that we do at the firm, along with how we think about development. So that's things like our leadership development model, what our evaluation process looks like. How we think the way we work and truly designing work to be developed, I know. So it's a little different than some of my peers who are a little more focused on learning roles, but I'm finding even them, their roles are evolving over time. I have been in the people space at McKenzie for the last 20 years. Started as a consultant. I've been here very long time. For those of you who don't know, a McKenzie company is a global management consulting firm. We are located in 68 countries around the world. We have about 40,000 employees and are considered a true accelerator for your development. And so we've been ranked, I think it's two or maybe three times. Now by Time Magazine, as the number one place for future leaders, you could imagine the bar for me in terms of how you think about developing lead leaders is pretty darn high. Maybe there's a few other things about me. So I currently reside on the West Coast. I have two boys who have my last one who's just graduated from college in like two weeks. Very excited and they are both going to be gainfully employed, which I am so very grateful for. That's amazing. We typically ask a question, why do people should listen to you? And I feel like after that intro, maybe I shouldn't ask it, but for the sake of fairness, I'm going to go ahead and ask it anyway. Heather, why do you think people should listen to you on this topic for today? A couple of things. I think one of them is I'm in an organization that is known globally around the world for developing people. It's about a leadership factory. And so thinking about how we develop books, I think, out of that would be the number one reason. I think the second reason would just be that what we're doing at McKinsey is really kind of at the cutting or bleaching edge of I think where the future of L&D is going. And we for sure haven't figured it all out. But I think we've got some pretty exciting things that we're doing here, as you all are thinking about, you're learning in development organizations, how you evolve, how you develop people, and explicitly not using the word how you train people, how you develop people in this age of AI with a constant change that's around us. Hopefully there'll be some things that you can take away and say, wow, I'm going to try this in my own organization. So I think that's actually a perfect lead in, because you corrected, you're not talking about training, you're talking about development. And we, at Retrored, we've tried really hard to move even away from learning to more employee development because it's about developing. It's not just about acquiring knowledge, it's actually about developing people. And McKinsey, maybe there's a finer point on it because your people are your product. It's not like you're having people do something to create a product, they're actually your product. If your people aren't right, then it affects all kinds of things in your organization. You started speaking differently about development in your organization. And I know this because we keep pretty close to apps on you all and kind of some of the things that you're doing. And in fact, your title isn't even, you said it wasn't a CLO, it's the Chief Learning and Development Officer. Development is in there firmly. And so I want to start with language. You are speaking differently. Why did you think that was important? And how does your team operate different because you're speaking about learning differently? - I mean, we're speaking about learning differently because again, we know that how do people develop, right? Really, how do they develop? And so much of it happens in the flow of work. And so as you're thinking as your organization, if what you're really trying to do is to upskill your organization in a particular area, build a set of new capabilities, catch people up on a set of capabilities. What you're really looking at is not about how many learning hours people go to, but really how they've developed, right? And there's just been tons and tons of research that's been done over time in terms of where people actually really did and that's in the slow of. And so at McKinsey, we really think about designing what I call a development ecosystem in terms of what are the full set of things that support that development. And one of them is learning problems. But there's a bunch of other things in terms of creating dress versus your own expectations and sort of how we think about evaluating you and providing you the information about how you're doing about those expectations. It's how we think about staffing and intentionally staffing people to studies to help them develop and grow. It's how we think about the actual, what happens on the ground of this study, this is apprenticeship. Then of course, there's, yes, there's learning which happens in the flow of work and outside the flow of work. And it's really thinking about that ecosystem. And I think what it means for me and for my organization is we actually spend a lot of time on those other areas, not just on learning. We think about those other sort of aspects. We-- a function broadly own our leadership development model. We as a function also own how we think about sort of design the work itself to be developmental. We spend a lot of time supporting a mentorship that happens in the flow of work and what that looks like. And so that's a bit of like how we think about our roles differently than I would say maybe a more traditional learning organization. So I think it's interesting because you're really talking about work as being the platform for learning in many ways. And I think you've said that you think the CLL role needs to evolve to something closer to a chief work officer focused on things like experience design, systems, integration and workflow enablement. So is that how your how Mackenzie operates today? And if so, when do you all start to make that sort of chip? So I would say a couple of things. One is Mackenzie has always been an organization where you kind of develop in the flow of work. We've always had a bit of an ecosystem around it. I would say since COVID we've actually made much more of a pivot to say that the learning learning organization is really a learning and development organization. And part of the reason we made that pivot is that what we saw that you're happened during COVID was kind of some of the the stuff that was just happening, the apprenticeship that was just happening got a bit eroded during the COVID time. And we said, you know what? It doesn't work for us to just hope it happens. Expect for it to happen. We actually talked a lot about what we call purposeful apprenticeship where we actually need to be purposeful about apprenticeship. And we actually did teach the skills of being a learner and then you teach the skills of being a teacher. And we need to think about what are some rituals and practices that we can embed in the teams to actually support and ensure that apprenticeship happens. And so that's when sort of the learning organization said, you know, we need to take a little bit of a broader remit and think about kind of development and not just assume the development is happening in the work because it sort of always had, but can we get better about it? I mean, I always say to everybody, you know, we've all had great teachers and not as great teachers. And we all know how much more we learn from a great teacher. What if an organization like the organization was full of great teachers all over the place? Think of what a strategic advantage that would be for that organization in terms of any time you get to sort of upscale at scale. That makes sense. So then if you're trying to maybe guide others towards this vision of what a chief work officer would be responsible for, but with that entailed that a traditional CLO does currently that job description does not include. Yeah, I don't know that I love the chief work officer because it gets into more of like design of the jobs and so I don't know. We're struggling with what the right term is. And I really like the word development because you can't be the chief development officer because that also suggests business. So I don't know what the right answer is. But one of the things I would really say is to them is how much of you're really thinking about what happens on your teams on the ground to actually support the development. And so there's a couple places that you could lean in. One place that I've talked to other CLOs about leaning in would be really thinking about are there sets of rituals that we can sort of support to embed in teams and so start to track them actually having with which would support that development, whether it's a set of feedback rituals or it's you know kind of more regular performance check-ins or it's providing more transparency on criteria for people to develop where they themselves are taking their manager. Here's the skills I have. Here's the skills I'm trying to develop to have one of those conversations. So what are some of the rituals that you can think about doing? The other thing I would just say is like can you start to think about teaching the skills to be teachers and learners? Those are two famous that you can do in your organizations relatively easily. I would say the second one maybe is more in your real house because you are a learning organization. But what if you were to teach some of those skills in your organization? What would that do to change kind of how what happens kind of in the flow of work? There's a bunch of other things you can do in terms of like evaluation criteria and what that looks like. We actually measure a apprenticeship at the firm. So I know who's got great teachers and who doesn't have great teachers. And I'm able to tell you that if you have great teachers, you know, you're two and a half times more likely to be advanced or you're rated highly because you have those great teachers. So we've done a lot around measurement. I feel like those are maybe next level thing that you forget to. I find this really interesting for a couple of reasons. First of all, I think I'm hearing a lot of sort of systemic thinking. So instead of like just change the person, like teach the person the things they need to know. You're talking about systems. I love the word that you're using rituals. Like how do we build the systems in the organization to make sure that this development happens as part of whatever else is happening? I love that. Interestingly, I've been crunching data this week and we asked 220 learning leaders kind of what the skills they think that they're going to need in the future. And for the first time ever, workflow engineering popped. So they said it in many different ways, but workflow engineering sort of became a thing where people are like, okay, if we can't do things the way that we've always done them, then we need to think about how we actually engineer them into the work that's happening in order to make sure that these systems work. I think it's fascinating. You've said that a lot of the learning happens on the job. Like even if you take the 70, 2010 model at face value, which I don't think anyone does anymore, but if you do, 70 to 80% of that learning actually happens on the job. It doesn't happen through sort of the formal courses. You have also made the point that the majority of time that L&D folks spend is on creating that 10 to 15% of the stuff and why that doesn't balance out. And so I'm curious, in McKinsey, as you're thinking through sort of this learning in the flow of work or going to where the work is or defaulting to the work first for development, what does that actually look like? Yeah. One of the other things that we haven't talked about in terms of like supporting development in the flow work is actually designing the work to be a bit developmental. I think Danny that goes to your point about like work engineering with the capabilities of AI and the fact that in a lot of places were actually identifying the workflow, the work itself, right? Whether it's in a call center where we're creating agents to support answering questions for a call center agent or if it's in a sales organization, where we're providing them, a gentick support for their workflow. There's a whole bunch of places where we can point to where people are actually identifying those workflows. If those tools are designed right, they can be both productivity tools and developmental tools because they're designed in such a way that maybe asks a set of questions, allows somebody to role player practice a conversation, gives them real-time feedback in the flow and work of how well what they said landed with or you know, for a call center rep, maybe they listen to 10 of your phone calls and then they give you a series of feedback at the end of those 10 phone calls, right? So there's all these ways that we can actually design in more development in a way that we've never been able to do before because in the past, without those agentec workflows, you might be reliant on the manager to sort of be designing in some of that stuff to be developmental and may have worked, may not have worked, but this notion of as we're redesigning these workflows to be more agentec, how do we really think about designing development in which by the way also potentially helps us with some of the cognitive offloading and the risk of kind of early 10 year folks not really being able to develop because if these tools are also supporting the development of whether it's a sales representative or also a person or a consultant, what does that look like? So I think and so when you say like what does that mean for your organization and how we think about it, I've got learning folks, learning designers now embedded into those workflow teams, right? So like the team that's redesigning like how we do, we call it client activation, we've got learning people a part of that team. The team that's designing a tool to support McKinsey seven step problem solving process, I've got a learning person, a part of that team. It's very different and fascinating. I love, love, love this idea. I talked to Lisa Christensen who works for you about this a little bit ago and she was talking about this very thing like how do you redesign the work so that learning just naturally happens and the thing that I find interesting about this is you're redesigning the work and you have access to that work to redesign which means you have credibility in the organization, but the other thing that I find fascinating about that is the impact that it has to the types of technology that we use for development. So do we actually need learning systems to the extent that we're currently using them now or can we just adopt, just like we adopt systems and processes and rituals that exist in the organization? Can we also adopt technology, especially some of these newer technologies rather than investing separately? I won't love your thoughts on that. Yeah, it's a great question. So I think a couple things. So one is, I think at McKinsey we feel very strongly that we will continue to have a really solid core in-person learning curriculum because we know that that is a extreme value for creating a culture, the connectivity, the community of McKinsey, which the benefits go way beyond any sort of capability building. And when we'd ever try to pull back from that in the past, we haven't liked the results that it has on the culture in our people. So we will have that. So that will mean we will always need some place to enroll the people. Right, the capsule of forces. So I think there will still be some set of learning tools that will support that. Maybe that's an LMS or whatever it may be. I think there's a lot of question about the rest though. Lisa and I who mentioned have been working a lot with the CLL lift group, which is a group of CLLs that are really trying to tackle some of the future business problems. One of the things is like, what is the future technology stack for LND? It's just a question that we're all sort of wrestling with. And so we've been sort of problem solving this. And the super interesting thing is as we were problem solving it, what we discovered is it was not an LND stack. It was actually the agentic stack for the future tech ecosystem. That was what it was. And it turns out that that same tech that is going to support learning. Also future interesting thing, interesting thing. And so we've sort of been kind of problem solving it. I've actually at Google Max right now because one of the things is a learning practitioner that is totally new and different is like, I spend a lot of time in technology. So I'm trying to learn. And they've been sharing some of the things on their future of what they would call the tech stack. And I was like, oh my God, it's not exactly, but it's dark close to what we were saying in learning. And I was like, oh, I'm so proud of us that we came up with Google. The reality is that is going to be the future tech stuff. Yeah. I'm also on the road at a different vendors event. And so they've been showing us analysts what they're developing, particularly their agents and all the like. And we, a comment I made to them was I said, well, how are you going to make sure that as you're identifying all these things in this HM, which is where I am, that one, people are maintaining the context they need to ask questions of the AI. So like, we can't. Otherwise, our expertise is going to stop now, right? Because we used to do all this stuff by hand. And then if there's machines do it, at some point, we're going to stop knowing how to ask the right questions of the machines. So how do we make sure we have the context that we have the learning as the agents go along? And how do we make sure that we're bringing people along on the journey so they're asking those right questions? So to your point, like embedding questions into the agents, et cetera. I think this is an opportunity for these companies' clients to really lean in and encourage their vendors to think about this, because it's not on their radar. They're like, what are you talking about? Absolutely. Absolutely. And every technology vendor I talk to, I say. So are you designing these new agentic systems to not just deliver productivity, but development? To really think about how you, as we get say, to our friends at Anthropic and Cloud and all the amazing things that they're creating, could you build that product in the same way, similar way, but it actually supports development alongside. Now, some of them are creating learning modes where you kind of turn on the learning mode. But to me, that's not quite the same as it happening as a part of when you're doing the work itself. Yeah, I'm certainly. I want to talk a little bit about technology and learning and performance. So in the study that Danny referenced, she found that the technology that enables performance with actually the number one thing that LND leaders are planning on investing in, that's from our state of learning technologies, our state of learning study, excuse me. I'm very impressed, Deja. Hi, Danny, you know what I try. But from the tech side, some of the things that are happening before and after the actual learning event are getting a lot more important and a lot more attention. So I'm wondering, for your opinion, as these LND and talent functions move in this direction, it's usually obvious that the metrics have to change. And so as you've been thinking about those, whether it's time to proficiency as a measure or maybe the amount of development that happens in the flow or work or whatever, what are you seeing? How do you think we should be measuring some of these changes? Yeah, well, I think time to proficiency, personally, is a great measure for us. So we say, what's the time to proficiency for a new hire to become kind of an independent consultant? That's a great measure for us. I think we also need to sort of then pick like what are some of the core skills that we really care about and say, how are we progressing on those core skills? So we've got a bunch of metrics around AI skills, no surprise that we're measuring. And it's both like, are they being credentialed and what evidence do we see that their proficiency level is increasing? So I think it will be maybe both a role, perspective like in this role, I'm ready to do this role, plus some skill metrics for organizations. I think that will be a lot of where we pivot to, which again, to your point, Stacy, means that I can no longer measure those things just inside my learning organization. It's a different way to think about people, analytics, and measurement skills progression. And I think I remember seeing an article you wrote a reference also to things like workforce planning, strategic workforce planning. And so I think about this, it's like, it's also understanding the granularity of where people are with their skills. So like if I'm trying to move someone from, let's say that whatever, this part of 30% of their job went away and they've got these skills that are, for the remaining 70% of their job, what level of proficiency do they have? And what other skills can we add on to them to then get them to that proficiency? And so that's gonna be different for somebody with one set of skills versus somebody with the different set of skills and proficiency. And starting to get a little bit more granular, 'cause we can't do really good workforce planning without actually understanding how long it takes us to get somebody somewhere. That's right. So I have two questions. One has to do with the skills of your L&D organization. So traditionally instructional design and facilitation and all of these types of skills were incredibly important to the L&D function. When you think about the team moving forward, what skills are particularly important to you? Yeah. So maybe we just start with the conversation we've just been have about performance development, right? So there's no shouldn't have been able to like a monitor performance, right? Monitor, you know, let's say the performance of our business analyst and their skills, right? How are we doing on building communication skills and then what are the interventions that you might tweak, right? And in the future world, if you go way out there, you might say actually I'm tweaking the skill inside my agent that happens to be showing off the inside, their outlook, whatever. And so what you're actually changing is not the faculty guide or the learning program, but it's actually the agent skill itself. I'm making this up a little bit, but like you can sort of see that vision, right? And so this ability to sort of understand data and analyze metrics and figure out like how are you gonna take a data source on someone's communication skills and the learning program and a manager's view of their skills that they assess and they do a self assessment that they create and take all the information and say, now what can I now know about these learners based on those skills? So that's one piece of it. A second piece of it is really the ability to build, you know, agente development solutions, right? And so the way we're thinking about it right now is that the learning organization might own like the building of the role plate agent and then the reviewer agent and like you can imagine, there's like six learning use cases and then we would have a coach and then we would apply those things into the context that that person is operating in, right? So if it's again, a agente workflow that's helping you prepare for a meeting inside that agente workflow that's helping you prepare for a meeting, I might have one button that says, do you want help preparing to the meeting and I get a coaching experience? And I might have another button that says, do you want to practice and then that pulls up a role play? But I'm not necessarily designing that those things just for that and designing them in concept and then they're going to show up in multiple places. So I need learning folks that know how to do that. I need learning folks that can design agente skills. Like what is the meeting planning agente skill for McKinsey? Very different. I need them like where everybody's learning cursor, which is, you know, to build stuff in cursor, they're going to need to be able to design rapidly, feel time out of content using, you know, agente tools, et cetera. So there's a whole bunch of different skills. Now I do think that kind of the core skill of like design and like that design expertise absolutely carries, right? Because it's, you're designing a development experience. You might not be designing a learning program, but you're just thinking what that development experience looks like. All the science around kind of what, how do you support development? Because again, we need to bring the expertise to these tech, you know, tech teams in terms of when I say, design it to be a developmental, what does that mean? Like what does that look like? And how do you do it away that we've been talking all about? That doesn't become like clippy. That is actually an, and the thing that's really interesting and this is what we're sort of figuring it out is, this meeting planning example, which is a great one. So the team that was building this agente workflow, basically built this thing that pulled together all the information from all your sources to help you prepare for meeting. It's now got all this information and people are like, wow, this is a lot of information's great, but now what do I do? Right? And so us working with the team to say, oh, what do you know if we create this little button that says, help me prepare and ask them to send a question that is a learning moment, but is also helping them prepare. Nobody even realizes it's learning, but I'm teaching it. I love that and I love the point that you made. This is kind of where I was trying to get and I-- And you got there, which is like those skills don't go away. Instructionally design is still incredibly important. Design in general is still incredibly important. The skills of facilitation are still important. They're just applied in very different ways than they've traditionally been applied. Yeah, it was funny because Lisa, we were talking about, like, you know, used to write faculty guides. Now you're writing skills. It's the same skill. I mean, it's literally the same skill. Right. I love that. So when I ask you a question, can I go back to metrics? Because data metrics is my role here, so I'm always going back there. And that is, if L&D is not designing faculty guides or all these things, what is L&D itself getting measured on? If L&D becomes, I think Danny had turned it invisible, you know, there's always this question of like, how do you show your value? And I know like great organizations like McKinsey, you don't necessarily get asked for that. But a lot of people do. So how do you show the value? Yeah, it's a very question we've actually been debating with our CLO friends as well. So I decimulate, think there's something in that sort of capability proficiency levels. If as an organization, you are responsible for building clipp abilities on AI, what proficiency level of your organization currently at, and what proficiency level can you get to it six months? And that's what you're going to be measured on is, were you able to pivot the organization to help you get there for six months? Now you would say, let's not all L&D, there's a bunch of other players in there, which I think is absolutely true. But I think those are the kinds of things we're going to have to start measuring ourselves on. And then again, it may be, like we talk about you're designing these tools. And again, maybe a joint measurement with the folks designing the tools about the user experience, do people feel more confident in doing the work with the tool? I mean, you can imagine, again, a set of like, how successful is that agentech workflow in actually changing the work and supporting the people to achieve whatever the goal is from that agentech workflow? Again, it'll be like a single learning goal, but learning will be a key part of making sure that happens. And so I think those are some of the things that we may start to shift to in a learning organization. And again, for us, we will continue to have like this in-person learning portal, in the great experience, et cetera. And then there also may be broader metrics. I mean, McKinsey, you know, have an annual survey about people feeling supported in their development. It doesn't say in their learning, it says in their development. Right? And again, learning's a piece of that puzzle. We're not the whole piece, but we're important piece. When things are really like about that, is it focuses on the outcomes? What capabilities, to what extent have they been developed in the six months? Are people feeling like they're more supported or capable in doing the work? So I think that that is an important distinction that gets Ellen DeYat of the, you know, just tracking activities to the, are we actually making the impact on the business? I've had this conversation a little bit with a couple of people on your team, and I think we had this conversation at dinner as well. You're worried about people coming into the organization and not having the work to practice on in order to get good at the job and be successful. And that's where some of this stuff comes in that you've been talking about. I'm wondering how you and your team sort of identify those things that they absolutely need to know versus the things that maybe AI can just take and they don't actually need to know. - Yeah. So we've actually done a lot of thinking about this, and we've actually really thought about like, what are the skills that we believe are gonna make McKinsey distinctive in the future, right? And, you know, for example, there's a set of problem solving skills, systems thinking, metacognition, really great judgment, conceptual thinking, creativity that we believe are gonna be absolutely critical and distinctive in the future. And so then we work those skills back and say, "Okay, then what is it that our new hires need to be able "to do with those skills?" And we're doing a couple of things. So one of the things we're doing is we're embedding more deliberate practice of those skills into the learning programs that we have. And the really cool thing about, we've always had deliberate practice, but now we're doing more of it, and we can do more of it because we have AI that they can practice with, and they can turn in assignments and get feedback, and then the really cool additional benefit is that I can see where their skill level. So I can actually show you that, or now, in the course of my onboarding program, I am seeing people that have a 10% increase in their problem solving skills and a 20% increase in their communication skills. And I can do that because they're doing deliberate practice exercises, they're turning them in to get the practice, by the way, they're getting immediate feedback on that activity, and then they do it again, and I can see how much better they got on that activity, right? So there's a whole bunch of benefits to that deliberate practice and they're getting more at that. We're also then, as we think about, where we're leaning in right now as an organization, when I talked about these workflows, and really designing them to be developmental, we're also intentionally leaning into workflows that we think really matter. And so again, I described one of these workflows is the core problem solving process in the Kansas. And we have built a, it's all in section, it's a new tool that we've just recently rolled out that supports the problem solving process. And for us, it's like starts with creating a problem statement and building that problem statement, which is now being AI enabled, and then building an issue tree, and then doing analysis the issue and the research and then eventually producing PowerPoint slides that really still we don't need to use PowerPoint anymore. We can just tell the slides, pretty cool. But again, that particularly that like problem statement step is super important for like all the thinking that we want. And so really, how do we design into that tool that the tool maybe provides more of like, here's why I'm thinking about it this way. So it's more transparent with its own thinking and its own reasoning to help people understand. It supports you asking aesthetic questions. It's monitoring you and it actually triggers something. If it's seeing you maybe building issue trees in the wrong way a couple of times, it actually triggers something for you to say, hey, have you thought about this? And so again, we're focused on problem solving because again, we believe some of those skills are critical and then we're saying, okay, then that's the skill I want to embed more deliver practice around. That's the place then as we think about a generic workflows that I want to lean into thinking about them to be developmental. That's just some of the ways we're thinking about this, to support our early tenure colleagues. One other thing I will say is that we do also believe that on the reverse side, like because we're doing some of this work to be more developmental in the work, we actually think we could at protest accelerate development even faster. - I think it's a good place to maybe take a quick breather. So this is lightning around Heather. We ask you three questions that you may or may not be uncomfortable with to get to know you're just a little bit better. So I'll ask the first one, what do you think your super power is? - Well, this is something easy because my husband and my children have always told me, my super power is my ability to focus. And so I can basically be doing something for work or for pleasure or whatever else. And the house can be exploding around me and I'd keep on it. - That's awesome. - That's a really, really valuable one. You can't tell me, because I'm like, no, no, no, no, no. - You know, you're just like, no, but I'm focused. I'm getting it done. - That's amazing. - Speaking a little bit of your home life, where did you grow up? - So I actually grew up in Cleveland, Ohio. Actually a little sub-arth called Shaker Heights, Ohio. My husband and I actually both grew up there and graduated from Shaker Heights High School for any Clevelanders out there. And then the fun thing is, well, we raised our kids mostly in California. We actually ended up back in Ohio for then for high school. And so they both graduated from Shaker Heights. - That was so cool. - Which is really nice, very fun. - Did you and your husband go to school together? Like at the same time? - We did. - We did go to high school together, I was a nerd and he was a jock. And so didn't actually get together until after high school, but yeah. - That's awesome. - Okay, this is one that we think you're gonna have some insights on and that we personally can use. What is your favorite travel hack? - I think having. And so I have either a blue week, a black week, or a brown week. And then I have either blue shoes, black shoes, or brown shoes. So it's like one pair of shoes, the black kit. And it's like I've got it for summer. I got a black summer week. I got a black fall week. I got a black winter week. Same thing. I got a blue, blue, blue. And I got a brown, brown, brown. So. - Oh my gosh. That's intense. That goes clear back to even your shopping. It's not just the packing. It's like, - Well, yeah. Because that's easy. Okay, I'm in the summer. The last week I packed was a black week. So the next week I'm packing is a brown week. - I love that. - But we travel all the time. It's like you can't leave any energy to spend it out thinking about packing. You just. - Yeah. So you literally, you can take a section of your closet and like put it in your suitcase. - You got it. - That's so cool. - And we can get in some more. It's like one jacket, one pair. You know, we can also go, you know, we can go deeper if you want. And it's a little scary. So I walk a go there. But. - Maybe you can build a skill and send it over. - Well, you have one now, right, Danny? - I do. - I actually have a skill for my wardrobe. - Yes. - Do you know? - I do, yeah. It's like, all right. Here's my wardrobe. What do I need to do to flush it out for spring and summer? And here's my. - I want to use that. - I mean, to add the packing well on your scale. - No, I need to add the packing. And I'm embarrassed to say that given that my hair is up and I'm wearing a black t-shirt today. - Yes. - I'm so glad you're actually wearing this. - Okay. Well, let's turn back a little bit to this idea of AI in employee development. And so I believe that in some writing somewhere, you said, "It's not an autopilot. It's copilot about AI's role with an employee development." So, as you think about how organizations are approaching AI, what do you think they're getting wrong right now and when do you think they should be treating it as a copilot? Yeah. So, it's been talking a lot about this. I think there's a lot of emphasis in organizations about AI for productivity and how can AI improve the productivity of our organizations. I think we're also starting to see a bit more of a pivot now, our organization to say, how can AI help me unlock new business opportunities for those organizations? I would love for us to think about the third piece of the pillar, which is how can AI enable faster development? How can AI help us create faster learners? And so to me, that's kind of the third leg and the stool that I feel like, boom, we're just on the cusp of really people thinking about and you've heard me describe some examples already about where I truly believe that AI can help us further accelerate that development and sort of move away from, you know, there's often this trade off that, you know, I talked to a lot of folks in organizations about where they're saying, like, how do I make the case for time for learning? How do I make that case? How do I create that ROI? How do I define the value for time spent in learning? And I feel like a lot of times we have this, like, whatever, the parallel curve that's where I say, like, here's the frontier and the frontier is like, how much time you get for development versus your productivity? And there's like, just trade off that you're making, which is like, now I'm going to learning so I'm less productive in my job and all of us learning are making that case. And I claim that if we actually think about the future and really unlawfully agentech, there's no more trade off because work is learning and learning is work. And there's just no more trade off. And I'm no longer having to make a case for, you know, people being able to develop per se because they're actually developing in that workflow itself. I'm believing, you grew about an AI evaluation tool that you all developed that might be a good example here. Could you talk a little bit about what it was and why you see that as kind of being relevant to us? Yeah. So, again, Mackenzie is a people place. We are very focused on the development of our people and we take our evaluation processes very, very, very seriously. And there's a ton of our readers, partners, time that is spent on evaluations of our colleagues for a huge portion of our colleagues. They get evaluated sort of more formally twice a year. Those evaluations will include sort of regular feedback that's happened over the course of an engagement that we call our performance reviews. Plus, then when the evaluator or the DGL takes on that case, they actually will then interview a whole bunch of people and of course, that an evaluation process. And so what we were looking to do is again, to a design and a genetic workflow that basically both improves the quality of the reviews and feeds that up for you. So we have an AI tool now that we use in the evaluation and we've been really thoughtful about where it shows up to to make sure we're not eliminating the role of the human, but where it shows up right now is it helps to transcribe all of those conversations. It uploads transcriptions of conversations plus performance of reviews that's happened over the course of the year into the evaluation process. It then will automatically help you create a summary of kind of all those of that information to give evaluators a really good starting point. But it also does some things to help the evaluators get better. So one of the things it does is it has a bias flag in it. So it goes through all raw materials and it looks for areas of bias and it flags them to the evaluator and it says, "Hey evaluator, you know, we think this is a potential area of bias." Now, you know, in the past we used to have bias training for all of our evaluators. Who may or may not remember everything when they actually get into the evaluation process, right? And we actually, it was kind of mandatory and all of our evaluators had to go to bias training. But now we've got it sort of embedded into the tool in a way that like flags things. And again, you've learned from it, "Oh, look, when I asked that question in the interview, that might have caused some bias. I want to ask those questions again. I'll ask it in a different way or now what I'm thinking of." So that's just one example of how we're thinking about embedding, learning actually right into the flow. So I think it's interesting, both from the vendor side and from the organization side, I hear a ton about AI fluency right now. We need to get our people up to speed on AI and competency and things like that. But it feels like Mackenzie is taking maybe a more nuanced approach to that. It's not so much like, get in there, train them, teach them how to push the buttons and leverage it for their own personal work. You seem to be thinking about it a little bit more systemically. Do I have that right? Yeah, it's a bit of a pivot or shift in the conversation. But yeah, so we are really thinking about, again, and this is a part of the work design and changing the work. But how do we think about the changing of the work and then what it means? And so a great example is again, if I go back to our problem solving process, which is like the core of what we do, rather than just teaching the tools. So we have a tool called Lily that is kind of our own chat GBTA that's based on Gemini, that's our own chat GBTA. And then we have these cursor, repursor, and we have chat GBTA and we have all these other tools. Rather than teaching the tools, I'm teaching you to problem solve. I'm teaching you how to build an issue tree. I teach you to build the issue tree. I then teach you to build the issue tree with the tools. I then think about, okay, now I'm going to reflect on you, the human building, the issue tree, you, the tool building on the issue tree. And where can I then take that issue tree that the tool builds and make it even better? And so really thinking about that sort of both the teaching process, but also then what it means in terms of our expectations around problem solving. So we're actually thinking about a lot of this as how do our leadership skills evolve in addition to just teaching the kind of quote unquote technical AI skill, meaning can you prompt and how effectively you can evolve? Can you worry about people becoming reliant on it? Yes. So why this notion of really embedding new expectations around problem solving are really important, right? Because there's a huge burden on fact checking. And it's interesting because there was always a burden on fact checking because again, you know, as you think about somebody built on Excel model, maybe they screwed up, they froze the cell and the model is not appropriately updating. There's always a need to fact check. But I think we run the risk of trusting the AI too much and not realizing we still need to do that level of fact checking that's associated with that, right? And so it's really thinking about fact checking. I would say one of the other things that we're really, we do a lot is traceability of the evidence of the tools, right? So like everything you get out of our LLNs, whether it's the evaluation when I described or our own internal is the sources. So you can always go back to the sources to follow that trail. If I've got something that I've asked about some research that McKinsey has done, it gets me the source, I follow the source, I can go check the source data. The same thing with the evaluation is I'm summarizing thing, I have the sources in the value, where I can go click and see actually what was the quote was what then the evaluation tool created. So this notion of really building and traceability, I think is something else we're doing, but doesn't stop the ability to actually be teaching these skills. And again, as you think about the AI itself and building it to be a more developmental tool as it returns something, can it prompt you to ask the next set of questions that says like, what's the counter argument to this? What are you missing? There's ways we can help that individual and support them on this. I'm interested in this point in particular in parallel with this discussion about how we're redesigning work. To what extent do you think that there are hearts of the work that you are being or Danny had to do to develop? Just because that was just part of how the work got done, no LLM existed, whatever. But we're maybe not, if you think about the work that we're going to be doing for the next, I don't know, five or three to five years, or just not that essential for development. Are you thinking about what points, what things can we pull out? Aligning the columns on PowerPoint slide. Who wants to do that? You really get it aligning with column, how do I change the size of the font? Are making sure it's the font that matches the client's style tab, whatever. There's a whole bunch of things that we do believe you won't need to do anymore. Now, ensuring that the message is compelling, relative to the data that's being shown on the chart, or that I'm building a compelling story for the client and I understand their context as I'm building that story, absolutely critical. The whole pieces of work, I think, about when I was a business analyst 30 years ago on how much time I spent cleaning data. Oh, go ahead. Right? Nobody needs to spend that much time cleaning data. So yes, there's certainly aspects of it that we believe will really change fundamentally the work and actually our ability to even have even more. Yeah. So then as you're thinking about the learning and the learning you're designing for, it sounds like there's almost like a wonderful words in your mouth. So the purposeful abandonment of some of those that focus because now you can move another direction. Yeah. Where we just, we were just having conversations recently or PowerPoint skills and like so can we start teaching people how to align columns and PowerPoint and our onboarding? Could we use the time to put this up in our and what is the moment that teaching them like that to do those things, right? Why? I want to ask, this is not on our list Heather, so forgive me and if you don't want to answer, you don't have to. I'm very curious about this idea of authenticity. It came out as a is one of our big five mega trends this year, this need for authenticity. And that is both from like a data security standpoint, like people showing up as who they say they are. But also in especially relational work, like the work that McKinsey does, that need to be authentically human is important. And I'm wondering as you're thinking through like the skills that you keep versus maybe the PowerPoint skills that we don't keep, how are you building in the responsibility of the individual to be authentic, to be their true selves and come across as their true selves instead of setting like a GPT. - Yeah, GPT, that's a great question. We've actually sort of embedded much more into our curriculum, this concept of leading self. And can you lead from confidence? We have this concept of are you a self author or are you a survivor? And the types of behaviors that happen when you're a survivor and that you're leading through fear and you aren't your authentic self versus what you need to do to be a self author. And this notion of self authorship is big at McKinsey, we talk about make your own McKinsey and what that looks like, et cetera. And so I think again, as we've been thinking about how do we need to evolve our curriculum, we've really thought much more about this notion of leading the self. And particularly for our very most senior leaders in particular, we have actually kind of doubled the amount of curriculum we do for our senior partners, relative to sort of what we did two years ago. And a lot of that is a lot of curriculum around leading self because these leaders like CEOs or other executives are dealing with faster moving pace, more challenges, their client have more challenges. And so it's even more important to them to be grounded in leading myself before I can help to lead others. And so that is something we're thinking hard about is what does that show up and look like? And interestingly, we're doing a lot of that in person learning programs is really focused on some of those skills, which again sort of take that, small, we do a lot of small, group kind of discussion led by it kind of a count of facilitator around some of those topics. - So earlier you mentioned this idea of apprenticeships and how they need to change a little bit in this AI agent. I'd love to understand how you and McKinsey are thinking about the differences. - Yeah. So I think again, the advantage we have with AI is it can support the apprenticeship model. So a lot of the pushback when I talk to other leaders about apprenticeship and rolling it out in their organizations, they say, my managers are already overloaded, like asking them to now be more intentional about teaching, asking them to provide the role modeling and the scaffolding and all these skills we talk about there needed in teaching. Like they just don't have time. And even if you say, hey, look, what you're doing is you're creating more capacity for you and your team because you're actually building the skills of the people underneath you, they're still like short-term, I just can't pull it off. But one of the advantages with AI is again, is this notion of can it help you be a faster learner? It can be the apprentice, there's from apprentice support in a couple of different ways. So it can support the learner, right? The learner can actually be more intentional about what they're trying to learn. They can ask AI in the first go around for the feedback on the thing. So maybe it's bringing up times with a teacher. The same thing is AI can be kind of always present as a teacher. When you don't have the, you know, your manager around, you might be able to send your document directly to AI and say, and the AI is like the manager where you are first round and it gives you feedback. And it's interesting, some of our partners and so your partners actually create kind of individual twins of themselves. Where they actually upload all the materials that they, like, great documents, what they really like to look like, et cetera. And then they basically use that as a first for me or to give feedback to teens. Then when they get the document from the team, they can then sort of start with a conversation about, okay, what do we think the client's thinking right now? What are the challenges that we might want to overcome? How do we want to position this conversation in the meeting? Is there any sort of pre-conversation we might need to have an advance or set for a versus? Hey, on slide six, you know, the data you have there doesn't match the lead. You know, we can shift the conversation where the AI is providing the initial development. And so again, that's where I think there's a real advantage of AI into providing that sort of first line of mentorship or coaching, apprenticeship, et cetera. We're seeing in a lot of organizations they're finding real value from some of these AI coaches for that same reason. Heather, unfortunately, you have to start to wrap up. Thank you so much for spending so much time with us. One thing in wondering though is, as you look out, you talk to a lot of other L&D leaders, what is one move you think that they should be making right now given this moment where it? So I would say, but pick a friendly business or an important work slow process and see if you can get yourself into the design of the world. Where do you have credibility? You do it in one place, right? Because if you do it in one place and sort of prove out a different model, you will then sort of be able to spiral that and build the credibility. So like, what's one priority skill, one priority workflow that you might be able to bring together to drive impact your organization in a different way? So I would just say start somewhere. Love that answer. Second question is, where can people find more of your stuff? You can certainly search for mckinsey.com. I published quite a few articles on mckinsey.com. If you go to my LinkedIn page, you can see a bunch of other pod tasks that I've done that will link you to those podcasts that they want to learn more. Great. And then we're going to wrap up with the final question that we ask. Every single guest who comes on our podcast and we call it the purpose question. And it's, why do you do the work that you do? I am platform about developing people and I am a platform to help develop future leaders, not just from my organizations, but organization, but for all of yours. And I take that very, very seriously. And I can't imagine a more important purpose. Heather, thank you. Thanks for listening. If this podcast got you thinking differently about how work happens, that's what we're here for. We dig into these topics every week in our research and in the Red Thread community. The Red Thread community is a place where people leaders and practitioners come together to compare notes, share what's working, and maybe challenge those things that aren't working so well. You can join for free at redthreadresearch.com/membership or if you or your organization are ready to go deeper, take a look at our professional and our enterprise memberships. For professional tier, for listeners, you can use the promotional code Stories, STORIS, for a 10% discount. And if you're a tech provider trying to make sense of this space, our tech consortium is where smart vendors are coming together to learn from each other and from the market. You'll find everything at redthreadresearch.com. I'm Danny Johnson. Thanks for listening and for caring about making work better. [MUSIC PLAYING] OK, that's a wrap for this episode of Workplace Stories. If you like what you heard, please subscribe to our podcast so that you can be the first to know about future episodes. For exclusive access to our research, insights, and a vibrant network of human capital leaders, consider joining the Red Thread Research Community. Visit redthreadresearch.com to learn more and to take your workplace to the next level.