Can AI make you a better manager? with Hilary Gridley

Hey, everyone. Molly here. When I agreed to come on as the new host of WorkLife, I was terrified. But I was also excited to learn from our amazing guests and to share insights that could help all of you, our listeners. One of my favorite takeaways so far came from Kaya Henderson, an iconic leader who told me that leadership is really all about listening and co-creating success with your community. So, I have a request for you. Please fill out the WorkLife Audience Survey. It's quick, it's easy, and it's linked in the description for this week's episode. And it'll help us all make this show the best it can be. Thanks for listening to WorkLife. If you are doing really thoughtful work, the AI can be an amazing amplifier for getting that thoughtfulness out at a bigger scale. But if you're just tapping a button and going in and saying, "Oh, you know, we need a Q3 strategy." Like, "Hey, Claude, write a Q3 strategy for me." That's what's going to be slopped. I have one pretty strongly held belief when it comes to AI and all the stuff that's happening right now. That belief is that there are things about being human that aren't going anywhere. The messy, relational, judgment-heavy stuff, the art and taste and intangible stuff, the part of work and creativity that is almost impossible to teach or explain. That stuff, not going anywhere. My hope for AI is that it replaces all the like repetitive, menial work that humans should never have been doing in the first place. You know, like data entry or repetitive rule-based tasks or, you know, complex research and analysis that can take months and months and months. Things that aren't actually a good use of the power or the scope of a human brain. That's what I think AI should be for. And if it can take over that work for us, then I think it lets us, the human, focus on things that we're uniquely good at. I've told you that once in a while I'm going to do an AI curiosity episode where I talk to someone who is deeply embedded in the AI world about a topic. And my goal with these episodes is to help us all learn what is real and what is not in all this AI chatter and also what is worth our energy day-to-day. Today, I want to explore something deeply human at work. Management. Managing people is relational. It requires judgment, taste, knowing and seeing the person sitting across from you. The best managers I've seen do something I've never believed you can systematize. They help people become better versions of themselves. To be clear, there are a lot of bad managers out there. In fact, bad management is something you are almost guaranteed to experience in your work life. But in general, even though many of us are still learning how to be graded at, management is, I would say, the human kind of work that AI will never replace. Because we're talking about two complex human beings and a lot of nuance and feelings and judgment. And yet there are people who are making the case that AI should have a really big role in management. People I respect and whose judgment I trust. Today, I'll be speaking with one of them. And I'm going to bring all my curiosity and all my skepticism. Can AI actually make you a better manager? Or does it just give you a faster way to do the job badly? I'm Molly Graham and this is Work Life, where we untangle the messy human side of work. Hillary Gridley has spent years leading teams at companies like Woop. More recently, she's known for her work helping managers use AI to coach people and give feedback and develop talent more effectively. In the process, she's found herself wrestling with questions that every manager eventually runs into. What is good judgment? Where does it come from? And is it something that you can teach? And can you teach it to AI? Hillary thinks that the answers to all these questions are more complicated than I've been allowing for. So today, we're going to get her take on AI and management. But we're also going to talk about things like judgment and what might have become more valuable, not less, as machines get better at doing our jobs. Hillary Gridley, welcome to Work Life. I'm so happy you're here. Oh, I'm so excited to be here. Can I tell you why I'm so excited to be here? I'm going to tell you. So your writing has obviously been so influential to managers working in high growth companies dealing with a lot of change. And so I've been a fan for a long time. But it wasn't until I read your water line piece, or I was like, oh my god, Molly was a wilderness instructor. And I was a scuba instructor. And I feel like these are some patico because back in the day, like my friends were going to get internships. And I was like, I'm going to go teach scuba diving. And everyone was like, that's a stupid idea. And I believe to this day that that has informed so much of my own management style. And I'm like very much of the belief that the future belongs to people who take inspiration from unexpected places because all the unexpected places are kind of getting mined. So I was like, when I saw that about you, I was like, I have a new theory now that like building this instruction is the best management training that exists. Totally. So I literally think that Noles, which was where I taught, you know, where I led wilderness trips, was my absolute best management training. I feel like so much of my like grounding as a manager came from their weight. So where, where did you instruct scuba? I taught scuba diving for a company called Broadreach. Great, great company in the Caribbean. We were leading adventure trips for young people. And yeah, I tell people all the time, like keeping a group of teenagers alive, a hundred feet below the surface of the ocean is like, oh, and now I have to like get a report approved to move forward. Like those are not on the same level of difficulty. Yeah, I feel like I feel like for me having a little bit of life and death, like not a lot of life and death because obviously like everything we were doing was pretty safe, made everything more real. Where like when you're in an office, you're kind of like, what's the worst thing that could happen? Do you know, it's like not that bad. Totally. I had a flow chart that I saw reference sometimes that was like, should you panic? And the first break was, is anybody in physical danger? Totally. And if it was like, no, it was like, is there a chance of imminent physical danger? No. Okay, like there's nothing to stress about. Like what are we worried about here? So it's such a good test. Yeah, well for what I would call the front country, which is like what most people think of as management because like if the answer to both those questions is no, then like, what are we stressed about, you know? Are you believe you probably shouldn't panic even if the answer is yes, but at least you are having an understanding of the stakes and panic after. No, I love that. Well, you're actually getting into, I mean obviously, like what I want to talk to you about today, which is like management, which is one of the most human things that I think we do and this whole world of AI that's showed up and you've obviously spent a lot of time speaking and teaching and writing about the intersection, like how can I help managers and people leaders become better at their jobs? But I actually want to start just by grounding us in something more fundamental, which is sort of like what do you believe? What is Hillary's version of what the role of a manager is? Sure. So fundamentally, I think a manager is just somebody who makes their team more than the sum of their parts, right? Like you are a multiplier for the team. They are able to accomplish things that they would not be able to accomplish without you. And what that means, how you do that, obviously, there's a lot of nuance there and it can change depending on the team. But I think kind of the fundamental piece is can you articulate what good looks like, right? And that's good in the role, that's good for the work. Like do you have a clear point of view for what success means for what good looks like? Can you show that articulate that to your team and can you help them reach that bar? Yeah, that's a great definition. I always say to that like, I think great managers make people the best version of themselves, make space for people to be the best version of themselves. But I love what you said too, which is like that one plus one equals four, right? That like you're not just every individual human sort of doing their thing. It's about how you make a team work together, which is part of what I hear in your answer. Okay, well, so out of curiosity has AI and sort of like the advent of these robots that we have a relationship with change anything about that definition for you? About the definition? No, so much of what I find very interesting about this. And I should say upfront, I'm very into AI. I use a lot of AI. I am what they call AI PILT, which I think it may be surprises people because I also consider myself a very human centered in how I think about work, how I think about management. And I don't see these things as at odds. And so much of my work with AI, both as a manager who was using a lot of AI and how I led my team when when I was at Woop. the wearables company, but also in how I build products and building AI products like working on our AI coach at Woop. You go through this sort of phase when you first start working on it where you're like, oh my gosh, everything is changing so much. Like my job every year was unrecognizable compared to the previous year in the years since Chatsubit came out. And yet you come out the other side of that thinking the fundamentals are more true than ever and more unchanged than ever. So an example of that is you hear a lot of people who are thinking about how to implement AI in the context of an enterprise, focused on context, right? We need to find a way to give AI the robots, as I like to call it, the right context about our company, about our strategy, about our people, our products, so that it can make good decisions. Now in theory you should be able to just write all of this up in a series of documents and load that into whatever system you're using and the robots can draw on that. But what happens when two teams disagree about what is the right context, right? About what is truth, which happens so often in companies. Like one team really believes that having a good user experience as measured by net promoter score is the most important thing. And another team believes that growth as measured by revenue per user is the most important thing. And those things are going to but heads constantly. And so you start to have disagreements about what matters the most, what is even the truth, what are we trying to do? So this exercise of trying to put all your context into a system so that the AI can use it is suddenly like, well hold on, do we who wins in a tie, right? Like what do we actually care about here? And what is actually the truth? And what do we consider data that is instructive versus data that is noise? And those are all deeply human and deeply managerial type questions that have always been the job of manager to solve that across the company. And now if you don't solve that, it gets so much worse because everyone can move so much faster and everyone can produce so much more. So just like the impact that one really high performer can have in a company has gone up a lot in the age of AI. So too has the impact that a person who is operating come from completely incorrect information can also have. And so I think of that about the again, the fundamentals of the manager's job to make sure everybody at the company, human and robot alike, has the right context in order to make the best possible decisions they can. But that's not a straightforward process. And it is as I said, more important than ever that managers are good at doing that. Yeah, there's so much inside what you just said that I want to unpack. And you're actually like really underscoring something that I was just talking to someone about yesterday, because there's this whole conversation in the AI world and the tech world as if almost like as if management's going away. Like you hear all these things around like the super IC, the super individual contributor, like everyone's job now is to do not to manage. And you know, you even hear like Facebook saying like, oh, now we have one manager for every 50 people, which implies like almost like management's less important or easier or something like that. And the interesting thing I think is that what I've actually experienced is the opposite, which is almost like every single person in the world that is going to decide to work with AI became a manager, because the skills of, you know, making a robot do what you want it to do are almost identical to the skills of making a human do what you want it to do, right? But I am curious like what your process was of sort of going from maybe hearing about these tools or being skeptical of them and certainly being skeptical of them as, you know, useful in management to actually like using them. Like what was that like for you? Yeah, so I was super skeptical. Basically all aspects of it, you know, can it do good work? Is it like anti-human in some way? Like all the reasons if people are skeptical, I was skeptical. And I started using it because in my job, I was working on building AI products, so working on our AI coach within the loop product. And I noticed that the engineers who were working on the AI product were working in a very different way than engineers and other part of the organization. They were working in what we, you know, now consider like a very kind of AI native way of building. And so it was a lot more of things like just sort of defining what they wanted the product to do or the feature to do and finding ways to evaluate whether it was doing that successfully rather than like writing the prescriptive code for exactly what it should do. And I was curious whether this was transferable outside of engineering. And so I started using it thinking about like ways I could do some of the things I saw them doing as a manager and thinking about, you know, in the same way that the engineers were sort of like less focused on the actual code writing and more focused on communicating what the code needed to do, were there ways I could do similar things with my team? And that's sort of what inspired some of my early experiments, some of which did not work at all, right? Like with any product, I, you know, I probably had, I think I made, you know, five or six GPT's and then one of them was extremely useful. And then I learned from that and I started making more useful things like that. And then I just sort of spiraled from there. And I realized that, you know, there's a lot of nuance in how you set up these systems. There are ways to design AI systems that I do think are kind of anti-human. That I do think sort of erode the cognitive abilities of the people who work in those systems. But there are also ways to set up systems that do the opposite that make the people smarter, that make the people better. You know, I've talked about how some of the tools that I've built are ways to help people on different teams who speak very different languages kind of communicate better with each other by by better understanding how that other team is like thinking about problems. And so that made me like very excited and passionate about learning more about where are the places that people are using this technology and ways that are helping people like their work more do better work rather than like more kind of scary ways. So, will you just give an example of like, you know, sort of your first eye open or some of your early eye opening moments when you sort of like use AI and like realized what it could do for you as a manager and a leader? Sure. So my first experience of doing this, I've always cared a lot about writing and communication and when I think about the thing I have been able to do best for teams, especially if I'm working with more technical people, I do often think it is on the communication coaching side. And I think communication is one of those things that a lot of people can be a little suspicious like oh I should spend a lot of time thinking about how to communicate it because that's somehow detracts from the merits of the work, right? As if it's an exercise in rhetoric and politics and influence, but so often it's just clarity, right? It's just being super clear and making sure that the person who needs that you need something from understands what you need from them can quickly, you know, think about is this a good use of their time and can quickly unblock you or give you, you know, the information that they have that you don't have or whatever it is. And so I would see this a lot when I was working with teams when they had to communicate with executives and especially CEOs where they would spend all this time working on some project and then, you know, they had to change plan and so they would need to go to the CEO and let them know that the plan was going to change and just get the CEO to feel comfortable with the direction that they were going to be going. And what would happen is oftentimes they wouldn't communicate it very well. And so I would spend a lot of time coaching my team on how to communicate in a way that prevented that from happening. But I didn't think it was the best use of my time to edit emails for people. And I also didn't think that it was a good experience for them to have to wait for me or like I didn't like this feeling that like everything has to go through me. So I set out to say, well, what would happen if I could encode all the things that I'm assessing when I'm sending email or doing this kind of executive communication? Could I encode that into a custom GPT so that anybody could upload something that they're working on some email they're going to send to an executive and get feedback and not just any feedback, right? Not just like what would happen if you put it straight into chat GPT and asked. But specifically, my Wait, this is work, so that were you doing that yourself manually or were you doing that? I just want to, like, for people that don't even know what a custom GPT is, like, yes, we just talked through your process because, by the way, this example is great, but just explain just a little bit more what you actually did. Yes, so I basically just made a document that had two columns in it, and I pasted a bunch of examples in the two columns based on my, whether I thought they were good or bad examples. Great. And then I uploaded that document into regular regular chat GPT, and I said, what are the difference between these two columns? Like, if you had to distill basically the things that I am doing when I'm editing people's work into clear criteria that I could use to evaluate work, what would those criteria would be? And then we chatted about it, and I said, you know, I like this, I don't like this. Let's try to get this to five criteria, and, you know, it's like clarity, it's action orientation, it's tone, it's things like that, and a very clear view into what does success for each of those look like? Like, what does passing look like? And write that out to find that. And then eventually I take that, and I turn that into a prompt, basically, which is just a set of instructions that an AI can follow. And I put that into a custom GPT, which is just a way to interact with something like chat GPT, but with a set of instructions that you've written, that it follows every time. And that you could share with your team, right? Exactly. Yes. So I'm telling it, assess this across the following five criteria, and give it a score, whether it is passing or failing. And here's what passing looks like in my own words. Like, this is my point of view that I'm encoding here. Did your team actually use it? Like, did it? Because like, I feel like we make these things, and we're like, "Everyone's going to love this." I mean, this happens without AI. Yeah, I mean, everyone's going to love this instruction manual. I just felt like, did you actually like deliver this to your team? And they were like, "Oh, thank you." Yes, but with some nuance, this was not my first, this was my first tool that was successful. Oh, okay. I had made other tools, and it's, you know, I tell this to my students too. Like, when you're building tools like this, when you're building tools that are meant to be an extension of how you offer coaching or support, it has to solve a problem for the person on your team. It can't solve your problem. And I think that's where a lot of things go wrong. If you make, if you're like, "Oh, my team's not strategic enough," I'm going to make a tool that coaches them on strategy, and then I'm going to tell them, "Hey, you have to use this tool to get strategy coached." Like, they're going to be like, buzz off, right? Both because you're inventing process for them, for no real reason, and also because it's just kind of a weird thing to do. So I, you know, I went to them, and I started with one person who I thought would be kind of friendly to this idea, and I was like, "Use this tool and just tell me if this is helpful," right? Like, no obligation, I just want to know if this is helpful. You know how annoying it is when you are doing work, and all of a sudden your day gets blown up by some, like, executive request. Before you respond to it, just use this and tell me if it helps you. And she came back immediately and was like, "This is so helpful, thank you." And so I think that's like, if you're thinking about how can I use AI in a way that helps my team, start from their perspective, like start from what are the things that are frustrating about their job? And this is just one example, right, which is a communication one, but there's so many. Like, start with anything that is a pain for them. And if you solve it, like, it's like product market fit for any product you're making, right? You have to have product market fit for the internal tools you're making for your team too. Because if you're trying to force people to use it, it's just like you're just fighting gravity. Yeah, you know, it's so, I mean, you're really validating some of the things I believe by what you're saying, because one of my beliefs, and we'll see if I'm right in like 10 years, is that the human side of this shit is just never going to change. Do you know what I mean? Like, it's still humans using tools. And I don't know, maybe someday we'll be managing teams of entirely robots, but it's very hard to imagine. So I just tell me, it sounds like you've had some experiences of it not working. Like, as you think about, like, you know, I loved your executive communication example, because that's a kind of pain that everyone feels. And it's like, where are these sort of tools and use cases best found other than just like pain from your team? Are there principles that you have when you think about like, what are the use cases where AI can really help me? Yes. So there's two that come to mind immediately for me. One of them is just where are places where the work going into it is not worth as much as what comes out of it? A heuristic I use is for whatever I'm doing, if getting 10 times better at that thing is going to make me 10 times more effective in my job, I actually don't want to automate that, because I want to get as good at that thing as possible. What's an example of that, like something where, yeah, what's an example of that? So I think one is a lot of people love to use AI for data analysis and synthesis. And I think some of that can be good, but I think synthesis of a lot of information is one of the most valuable processes that you can put your brain through in order to truly internalize whatever that information is, right? If you're a product manager, it is so different going and talking to a bunch of customers and then looking at all your notes and trying to kind of put buckets around what is the information that's relevant? What are the insights here? And especially what are the non-obvious insights here? That is such a different experience, even then reading a well-synthesized report, but so often the synthesized reports you get are not well-synthesized, right? They are lossy and you lose a lot of the really valuable information because the act of doing analysis, the act of making the report is only partially about the artifact at the end, so much of it is about like what it does to your brain to get to that point. And so if you are, for example, a product manager, your ability to have a conversation from which you pull out an insight and are able to sort out the noise from the actual like, you know, double click, this is super important, is one of the most important skills that you can have. This is true for many jobs, right? Anything where you have to make strategic decisions. And so if you start outsourcing that and saying, "Oh good news, I don't have to read customer interviews anymore," or "I don't even have to do them," because I have a robot that sends out these surveys, writes the survey, sends it, collects it, analyses it, and just feeds me reports, like you are robbing your own cognitive ability there. And so there are things like that where it's like, I think there are ways to use AI to extend your capabilities, which I can talk about, but I'm very cautious about it. Whereas there are plenty of other things like, you know, formatting slides is not something that I necessarily think you need to put as much time as I have lost in my life. I mean, yes and no, like there is value in getting an intuition for visual communication, but you don't need to put the same amount of work in every single time you do it. Totally. So do you have like an easy list or just like a couple of other examples of like something that actually feels very human, but that AI has actually been wildly helpful with, like just give me one or two other examples that surprised you like how effective AI was at helping you with it. So one thing I, you know, I've said that I'm careful about using it in situations where you're still low on the learning curve, but I actually think it can be tremendously helpful at teaching people how to do things. And especially at sort of handholding people through a process. And so an example of that is as a product person, a lot of what I'm thinking about is, you know, I have some hypothesis that our product would be better in this way, and maybe I'm right, maybe I'm not. And so I need to figure out how to validate or invalidate that hypothesis in the cheapest and fastest way possible. And when you work on a product team, a lot of people come to with their ideas, but there are many great ideas. How do you decide if this is better than any other idea? Well, the first thing you need to think about is like what assumptions am I making that would have to be true in order for this product to be successful, right? And then you can kind of isolate which of those are the riskiest assumptions. And then you can think about, all right, is there some way that we could design a small experiment that would test that and get us some data to help us see if this is actually useful, right? And maybe it's just making a prototype that we deploy to 40 people internally and just see how many of them end up actually using it. But that's a process that like in my head, I'm just kind of like going through that is maybe not obvious to everybody. And so I made again a custom GPT, which is just an, it's so easy. So it's so easy to make a custom GPT. You literally just write instructions in English, plain English, do this, do this, do this, and then the AI follows those instructions. So I made one that had the instructions for walking people through that, where people would input an example of a feature idea they came up with. And then the AI would say, OK, the first thing we're going to do is unpack the main assumptions that would need to be true for this feature to be successful. And then it would give a couple ideas, and it would sort of go back and forth with the person doing this to eventually lead them to rather than coming to me with a feature idea. They're coming to me with data, right? They're coming to me and saying, hey, I ran this experiment and here's what I learned. And that's the kind of thing that I can then take to like my team, my stakeholders, and say, hey, we need to move forward with this. That's a really cool example. So I want to ask you because, you know, listen, one of my beliefs is that there's probably more bad managers in the world than there are good managers. I don't know if you agree with that. OK. Yeah. But and part of what I hear you saying, as you're talking, is that to some extent, using these tools amplifies your strengths. Do you think there's a world in which AI is making better managers, or are we just kind of amplifying people that are already good and already not so good? Oh, good question. I think first, like, I talk about the, you know, this, in theory, the idea of, I'm going to take what's in my head and encode that into a custom GPT so that other people can use it. Technically, that's incredibly easy to do. And you're just typing sentences on a keyboard. Yeah. It's actually very hard to figure out, to get to that level of clarity in your own head around what is the process I follow here? What does good look like here? Like how do I think about what good work is and how we get there? And so I tell people all the time, like, even if you stop at the step of just trying to, you know, write out how you would walk somebody through a process to teach them how to do what you do, or stop at the step of write out the criteria for what good work looks like or what you care about, even just that would make you better than 99% of the managers out there. Like the, you get so much clarity in the process of trying to articulate what good looks like that makes you a good manager. And I feel like that for everything I've done, where I never used to think this much about, like, how would I explain how to do my job to another person? Or like, how would I explain what I think, you know, a good writing is, or what good product design is, or any of those things? And merely the fact that I am better positioned to explain those than I used to be. Like, in theory, that's me, right, that's still my point of view. But I never could have gotten there without the help of AI. Hmm. That's super interesting. So it's basically caused you to have to be, almost like, to be much better at making what is implicit, explicit, like that you, things that you're like, things that actually could be defined as like taste or judgment, like you're actually having to like write down criteria for, is that, is that actually what you would say? Definitely, and it's a challenge, and it's a fun challenge, and it is part of why I was initially very skeptical about AI, like when I would hear about these AI writing tools, I was like, never, never in a million years, what I touched those. For reasons including, I'm like, oh, you know, what I, what I do is so, it has such a genese qua, like I, I couldn't possibly like, you know, articulate the principles that if you can match those principles, you can do good work. And I feel like it has been a humbling experience for me to recognize that, like, no, you can, you can find ways to define good in anything that you do, and you should find ways to define good in anything you do. If you care about somebody who is like able to help other people do good work. And so it is, yeah, I think it's been a really interesting process in my own realization along the way of both the value of doing that type of work, and also the fact that it is like, yeah, if you, if it seems impossible for you to do, like if it seems impossible for you to explain to someone else what good looks like in your domain, in your work, like you probably haven't thought about it enough as a manager at least, because like, you should be able to get other people to that level. Yeah. You recently wrote about the fact there's been a big shift in the last year that managers that you work with and that you sort of survey have gone from saying that their biggest problem was getting their team to use AI. And now they're saying that the biggest problem is getting people to stop using AI badly. Will you just explain that for folks like that shift and sort of the trend that you're seeing? Yeah. So when I started teaching about AI and management, it was early 2025, and the number one question that I was getting was, how do I get my team to adopt these tools? And so a lot of the pain was, you know, people are resistant or they're busy or they're stuck in their ways, a million different reasons, but they just could not get people to use the tools. That's not really the case anymore. I mean, maybe in some places, but what I'm hearing way more is like, oh no, everyone started using these tools. But nobody was like, like everyone is using different tools, they're using them in different ways, and we're all drowning in slop docs. Like, I just get sent these 20 page documents that say absolutely nothing. And it's just like nobody's reading them. It's not clear who's writing them. And it's just creating chaos and a lot of companies right there. Like what do you think is sort of what I would call lazy AI use, but like what is, what do you think is creating all that what you called slop? But it's garbage and garbage out, right? Like, I like to say that the AI sort of, it shows you what's going on inside a person's brain, right? And it is an amplifier. And if you are doing really thoughtful work, the AI can be an amazing amplifier for, you know, getting that thoughtfulness out in a bigger scale. But if you're just like, tap in a button and going in and saying, oh, you know, we need a Q3 strategy, like, hey, Claude, write a Q3 strategy for me, that's what's going to be slop. Like to me, it's like, there is no thoughtful input. And so it's just taking nothing and then turning that into 20 pages. Yeah. Well, and I think you said somewhere that, because by the way, like this is, I think one of the most interesting and complicated things that's happening right now is everybody's like, I'm using AI and it's like, you're using AI to produce shitty results that actually make everybody else's job harder. You know, people are talking about attacks on the person that you're sending something to. Yeah. I think the other thing that worse that I'm seeing a lot that people are talking about inside the communities that I lead is, you give someone feedback on a document and they almost don't take accountability for it. They say, oh, but Claude wrote that or but Chatchee, Peter wrote that. So almost like as if they had a bad intern, you know what I mean? Yeah. But they're not taking accountability for their own work. So I'm curious, like you said somewhere that like, you can't solve this problem by teaching people tools. You have to teach people how to think. Like for all the managers out there that are struggling with this, what does that actually look like and is it, I mean, A, is it possible, but like how do you actually, how do you teach people how to think and how to sort of like, not produce slop? Yeah. So I think there, you know, there's kind of two approaches to this. One is you just hold the line on what the outcome is and like let people figure out how to get there. And the other one is you kind of hand hold people through the process of getting, understanding how to do that work. And I think both, both can work in different scenarios where one of them is the first one is sort of the more hard-ass like somebody puts something on your desk and you say like, this is not acceptable. And I will not read it until I read the first three sentences and I am not just like this is immediately obviously Chatchee, Peter and by the way on that one, like I think a lot of managers struggle because they say, oh, you know, I was pushing my team to use these tools and now they're using them and it feels like this awful monkey paw wish where I got what I wanted, but in a horrible sinister way. And then you feel like you can't chastise them for using the tools because you're the one who made them use the tools. And to that, I just say like the answer is just it's not done, right? It's not, hey, you did something wrong by using Claude. It's just, okay, great. You did the first step and now you have to do the work. And so I give that just like as people who are looking for kind of a script of things they can say, if you're ever like, how do I call someone out on using Claude? Like don't, don't necessarily tell them they were wrong to do it. Just be like, oh, this is a great start. But like for me to feel like this is acceptable work, like it needs to not feel AI generated because that signals low effort. So go ahead and take another pass on it. So that's one way. It's just like hold the line on the outcome. I think the other way is just you have to show people. And so I do a lot of demos. I think demoing, it's like, hey, it's great for you as the person doing the demo because it forces, again, it forces clarity in your own mind to get to a point where you can teach it to someone else. But it also is so, so, so helpful to see how other people use these tools. Because it's not intuitive to a lot of people. If you're somebody it's intuitive for, like that's great. And so you might struggle, we were like, why can't these other people just figure it out? But I think a lot of people really struggle with it. And so in team meetings, I will put up my computer and I will say, let me show you this problem that I was dealing with this week and how I solved it using AI, or where I brought AI in and just let people watch. And I hear time and time again that that is, my team loved it, I do that in my course, people love it. So I encourage you that, if there's something you know how to do well, just show people in one-on-ones, right? If someone comes to me with a problem, sometimes I would say, can I show you how I would approach trying to solve this problem with AI? And then like the kind of pair programming concept of letting people work together on things. So if one person on your team seems like they're really good at doing this, telling them like, hey, can you help this other person and walk them through how you might solve this problem? And really positioning it as like, this is a leadership opportunity for people who are motivated to get good at this, because you are just like, you know, giving heaps of admiration and reward to this behavior, and that's what is going to like motivate people to do more of it. Yeah, I like that a lot, because I think, you know, there's sort of a common management thing about like, you can punish the behavior that you don't like, but you much more effectively typically is like, rewarding the behavior that's actually producing the results that you want. And I think part of what I hear you saying is like, the more you can hold up good examples of like, this is great. Here's why it's great. Here's how it was created and help people see what good looks like. It's one of the themes here. The more they're going to then be able to think, okay, how do I, how do I do that? Okay, so let's just as a closing point, let's imagine I'm a manager that has maybe used AI a little like played around with it or used it to help some of my own work, but certainly not started to think about how to use it with my team. Like, where would you start or what would you, what would you, you push someone to think about in terms of like what, what use cases are a great way to get off the ground or just like what advice would you give them? So I think the most important thing is just you need to use AI a little bit every single day. And if you want a more structured way to do this, I have a free program that's called couch to 5k for AI and it's available for free at couch to 5k.ai. And it has structured exercises. They really take like two minutes, some of them take a little longer, and you do them every day for a month. And it takes you from kind of chatting with AI, asking questions to actually building with it and building things that help you do your job better, automating parts of your work, kind of having it as a chief of staff that just helps you do the best job that you possibly can, teaches you how to think about it, how to break down tasks, all of that. But you know, I think there are plenty of ways to do this. I think the most important thing is just that there is often like an activation energy hump of just like sitting down and doing it. And I think a lot of people think that they need to carve out a bunch of time or like take a course or whatever it is. And it's like just just sit down and try to do something that seems a little bit too hard every single day. Love it. Awesome. Well, Hillary, thank you so much. This was really fun. I really loved learning from you. Thank you. Thank you for having me. That was such an interesting conversation. It really reinforced some things that I already believe, but I learned a lot too. So like, one of the things that it really underscored for me is that so much of what you need to be good at to be a great manager and leader is what you need to work well with AI. To some extent that it amplifies things we're good at and we're bad at. In all my work, I would say one of the things that a lot of leaders are really bad at. And honestly, underinvest in is clarity and communication. We think that because it's in our head, everybody understands it. And it's so easy to forget that the people that work for you are somewhere else. They're on sometimes a different planet. Sometimes they're just on a different part of the same planet. But how do you bring people with you? How do you actually help them understand all the context that you have, all the decisions that have been made? That is hard regardless of AI. It's something that I spend huge amounts of time working with leaders on. And part of what I heard Hillary say that was so interesting is that working with AI can make you better at your job because you actually have to explain yourself. You have to take these things that are implicit in your head and make them explicit. Her example around what does good executive communication look like? That's something that I think a lot of people have experienced on both sides. I know how to send an email to my boss and get them to agree to something, but I see people fail at it all the time. Or if you're leading a team, I know how to convince the CEO or the board of something. But when my team does it, it's like, oh, but have you ever stopped to say, can I explain to someone? Can I teach them what good looks like? How do I actually explain that? Not just in feedback, right, on a specific presentation, but in principles, in actual criteria that say good looks like this. Like Hillary's point about it's not ABC in terms of grades, it's pass fail. Like, what does good look like? What's above the bar? Challenging yourself to teach others makes you have to be more clear. That's the use case to start with. And I love that. I hope it was helpful to you. We'll see you all next week. Work life is a production of Ted and Pushkin Industries. This episode was produced by Isaac Carter and Leah Rose. Band-Band Chang is our story editor. Mixing by Hans Dail Shee. Ted's executive producer is Daniella Ballerizo. Constanza Gallardo is the executive producer for Pushkin. Special thanks to Roxanne Highlash, Valentina Bohanini, Laney Lot, Tonsikas Singh Manivong, and Ashley Murphy. If you like the show and want more, come join the discussion on my sub-stack lessons. I'm Molly Graham. Thanks for listening.