AI4U | Episode 16: AI Native, and Who Still Gets to Decide

Welcome back to another exciting episode of AI for you. I'm Phil Mman joined as always by Jay Ta. We're both professors of analytics and AI here at the Fairfield Dolan School of Business. Jay is also the director of the AI and tech institute at Fairfield Dolan. We're joined today by someone very special, my sister Senya Mimman. She's my older sister. I've looked up to her for all my life. Uh I followed her to almost every school she went to. Um she uh got her bachelor's at Harvard uh in math and economics. Then later she went to get uh I think she was in the inaugural group for the masters of applied positive psychology under legendary Marty Seligman at UPEN. Then she went to get her PhD at Stanford in organizational behavior. So what I'm really excited to talk to you about today, we all are, is um AI native. What that means, where's it going, what what should we think about it, where are the pros, what are the cons. Um it's hard to imagine someone better placed than you to talk about this situation. You founded um Silicon Valley Change Executive Coaching about 20 years ago. You're you're a chief people officer in AI. um you have experience with this stuff and it's on everyone's forefront and mind because we're worried AI will replace us as human beings, but what happens when AI replaces us as as companies overall. So, welcome to the show. >> Great to be here. Hi, Philip. Hi, Jay. Great to be here. >> So, uh what is AI native? I guess we should start with that. >> What do you think AI native is? >> Um well, I think there's two uh there's miscommunication. One is sometimes people want AI natives, the plural. So when what that means is learning about AI and being fluent in it just like you might be a native in the internet right which is everybody nowadays but in 1999 it was maybe six people. So similarly with AI natives people who are comfortable using AI in every aspect of their life that would be AI natives but AI native uh when mentions a firm the way I think about it the best example and I don't know how successful the company has been but in my mind it's um Elon Musk's macro hard. He says it's supposed to be entirely AI. It decides who to hire, what strategy to do, who to fire, what to do, all the accounting, all the um business decisions, everything. I'm not even entirely sure where humans are involved. Maybe they're kind of oversee it, like you might have a board. Um but that to me, and I have no idea if they're working or doing any whatever, but that's the vision I see of AI native is um that the AI is really in charge and it uses humans as an input. Mhm. So you're seeing it both as a type of person, an AI native, like a person who speaks French, someone who just almost has grown up from with it, and you're seeing it as a type of way that a business can operate, which is AI fully functional with we don't know how, but some governance or some oversight. Let me check with Jay and then I'll I'll also give my definition. What would you say AI native is? >> I always see this as a spectrum bes beyond what Phil said. You know, one end of the spectrum is human. That's what we have been right and the other end is what Phil said you know AI native which is um AI running everything and maybe human oversee I think we should talk about human overseas part as well but my understanding of you know when you the AI native part is also a range on its own. So my thinking of the beginning of AI native the lower bond of AI native is you stop designing business or business processes for human you start designing them for AI. So AI as the first class citizen that's my understanding >> interesting so you would say AI native is almost like search engine optimization but for AI like let's let's design for AI. This website's going to be read by AI. Businesses are going to be conducted by AI. So you're kind of you're beyond what is one company doing for it and you're saying what if all companies were already there. Is that what you're saying? >> We have to start within the company you know from the organizational knowledge you know because currently we write these memos and everything to store organizational knowledge and institutional knowledge but the audience is always human. >> Yeah. >> But now we should have the second audience and we should promote that second audience actually to be the first audience. That's what I'm saying. >> I I really like what you're saying. when I've been operating as chief people officer or as fractional chief people officer I haven't had that mindset in mind probably because my stakeholders are the company making sure the company becomes AI native but I like what you're saying because it's sort of it's the purpose or it's the vision where we're going to go to so if we keep building for people will keep consuming but if we build for people and AI there's going to be more consumption so I absolutely can see where you're going >> so let me jump back to what both of you are saying >> I would give a definition and an example of AI native so AI AI native. If a company were AI native, then all of the main processes are done by AI. All of the main decisions are done by AI maybe with with some like an oversight or uh a process where things may go yes or no. But what's different about this kind of thing? So that's basically like macro hard the way you described it. But I think what's different about creating AI native whether you're a company who already exists or a company who is just coming into being formed is how do you start from that base? So if you're a company that's just starting right now in this year, amazing. You can go AI native. Everything you build, right? You're you're smiling because you can see it. Every every vendor you work with, every single process you start, everything can be AI native. What's trickier and most of the companies that we now know in the world, they are they didn't start in the past year, two years. So how do you take a company that's been successful, that's been working, and then make it AI native? And that's what I think AI native is the most right now which is a company has processes that are working. It has people that are doing a lot of things. How does it change those processes almost with having a new eye? So you've come into a situation what is your beginner's eye into looking at it. So if that were the case then the analogy that I would make if the definition is how do you look at a business completely from scratch like you're setting it up for the first time even if it's a 20-year company that's incredibly successful. The analogy is I went this summer I was in Italy with my kids and my at the time four-year-old daughter I put her in Italian day camp for 4 days. All they spoke is Italian. Some of the counselors spoke English. She doesn't speak Italian. >> But it is a different way for her to learn Italian if she's there than if I teach her or if I give her some app to learn. And I think of that analogy with AI native in business. How can a business imagine that they're already there in it as opposed to they're figuring things out or learning or implementing one process or another? >> So what what does that mean? You would create an environment for them where they like a camp where they have a portion of their businesses running in AI so they can see what it looks like. >> You're asking what would I recommend a business do if they want to be AI native? >> I would say you've got to have four things. Actually I've thought about this. So I would say it's it's almost like a game plan analogy. Uh like a sports game analogy. You have to have a game plan. what is and that has to come from the board or the CEO. We need to be whether it's AI native in this department or AI native throughout what what is that vision next is the sea level the executives need to be bought in with specific KPIs for their departments of getting there and then and this I think is actually critical and I've spoken with both of you about this there's got to be a core team of people who absolutely love AI and are willing to move it anywhere and move it forward I call that the huddle sort of the people that are it might be seven people we're not talking it has to be hundreds so now we've got the game plan the CEO board directive, we've got the executives, we've got the huddle or the core team, and then we have um grassroots or managers. So frontline people aren't using AI, I don't think you'll ever get to AI native. So my thinking is if you've got those four levels, you can implement start at any of those levels, push all of them, but then that's going to give you the the complete change around from something that you're learning Italian to you're actually speaking Italian. >> I love the four levels. I I think that's you know from my practical experience that's what I observed but I would argue on one point is you know you said you can start at any level but based on my observation or not only my observation MIT did a study last year on you know the top down and bottom up methods you know you start from the sea level the top executive then you know you have a commandment all the way down to the you know the workers or you collect the demands from the workers and you move all the way up to the strategy and and initiative level and they actually criticize on both. When you start top down, when you start making the strategy first, the top level people, you will be surprised. They don't really know what's going on because a lot of times it's it's not that they're that out of touch is sometimes you have to make um you know tactical decisions that's not viable. you have to make exceptions and and when you make strategy you don't really see those exceptions and deviations and things like that. So your strategy is actually out of touch when you get it, you know, druple down to the very bottom level. These people doing the actual work don't don't re really recognize what is asking of me because I never seen this thing before. This is not how I do this every day. Where you go the other direction you know um people start saying oh I use GBD I use cloud I use this this is wonderful for my work. then you cannot really generalize at the strategy level say what do I recommend here so they actually find out um over 80% of the initiatives and investment get squat because of these two ways I I would love to hear your opinion on when you say they can start at any level what do you think about that >> uh so what's the conclusion from what you've just described if you you shouldn't start at the top and you shouldn't start at the front line what's the conclusion So yeah, I was I was I was going to ask you, you know, what's your >> jump in if you want, I can jump in. >> No, no. I I think you know, I was going to ask you by your recommendation, but my recommendation to this is I always take a middle out way. The middle out way is start with the thing you call the huddle. Start with that. So you find the real processes, real pain points that you will actually impress, you know, certain people, your your future champions in a certain department. They'll see real tangible output of incorporating AI into their work. And then this I think you know when I say middle out, it goes both directions, right? I would recommend you go down first. You go down because the hurdle would be at the team lead department head level. Then you go down to the work and say by this thing I just built for our team you know see how it how it does. Once you you cover the department then you can go wide at the team lead level have more departments adopting that. Then once you have enough um departments adopting your your your solution or you know whatever that is your your your AI tech you formulate a strategy. >> Yeah, I see what you're saying. You're saying start with a huddle go down then go up. >> Yeah, >> I totally see that. >> Oh gosh. You're saying like sa decide what what's the best way. Um >> I have to disagree. >> Okay. >> I was thinking about it. So I I if there's not an initiative from the board or CEO level, all of it is going to fail because uh it's a matter of resources. Are there resources for the huddle for the main team that's going to build things? Are there resources for uh even time resources for frontline to invest in things? So you're right, I did say start at any point. I think what I really meant is start at any point as long as there's agreement that the company is going in this direction. So having said all that, I will tell you the biggest place where I do agree with you. If within your company, you don't have a core team of people that you think are amazing and that are solving some of the main problems within the company like how are we going to increase profit >> by using AI? How are we going to decrease cost? How are we going to increase efficiencies? If you don't have that core place, you're going to be u working around. You're basically like if you're we started joking that maybe we're making a recipe. You're basically like putting in the dough but you've got nothing in the middle if you don't have the core people because they're they're such an energizer. So people who are for for example something that I've done at companies where I've been chief people officer is I've set up an AI center of excellence and that center of excellence that that's the name for that huddle. Yeah. it. If those people aren't there, they're not energizing the frontline people in terms of what they're working on and they're not getting what the board and the CEO need most, which is data. This is data on how it's working. So, kind of I I would disagree in the sense of let's let's build all the levels in at some point, but I would 100% agree that if you don't have that huddle place, the key people, >> it's I think that just from a psychology point of view, the motivation starts to decrease. Can I ask you about the centers of excellence because we've all been involved at those in various corporate environments and it's usually just another meeting a week. Uh and and >> but can I just pause that is so true that is so true and so frustrating. >> Yeah. So how do you make an excellent center of excellence especially for AI where people want to build? >> Yeah. I would say do it small and be very limited in what you discuss. what what worked when I was creating a center of excellence in AI is everybody wanted to be in it actually it wasn't just the people who were incredible and were working on the company's biggest problems but their sponsors also thought that they should be in it and I can see that point of view and other people which who were related in terms of their work wanted to be part of it and that's a very important thing because you want that inclusivity so you share what's going on but I would say keep it small first and then I had two questions so for the first six meetings of our meeting and then opened it up. We actually brought in more people. We had just the core team. In our case, it was something like seven people. But I started every meeting with we are here to talk about what's working and not working. What are you building and in what you're building, what's working and not working and what metrics are you using. But I have to pause and say we were in learning phase. So if you're starting a center of excellence where you're in output phase and you're going straight to output, it might be different from that. You might start with these are the KPIs we need, but we were in we need to learn what's going on and what's working well. which is why I asked those two questions. >> The first time I'm in a meeting like that, I would enjoy it. But the second time and the third, it it it's a lot of repetition, is it not? >> It's not cuz they're building different things. So the people are building different things. The way we uh varied it from that point of view is some meetings we would just have cursory everybody. What are you building? What's working? What's not working? What metrics are you using? And we often went deep and that's what the people there appreciated most. So you're working on this project. walk us through what was the goal, what did you try, what didn't work, where did you have to go back to it and get 17 different permissions, what didn't work then, where did you not get buy in from your internal stakeholder who was going to use this thing. >> So they love the deeper dives because then what ended up happening and this I did not know this was going to happen. People would say you know what you just worked on can you and I just have a meeting next week because I'm working on something similar. >> That's that sounds that makes sense. Um but then the other part is there a way to use AI >> to make the AI center of excellences more excellent. >> Um the main ways we did it was planning for meetings and summarizing meetings. So I I don't know what do you guys think? Let's let's create our own AI center of excellence. What would we do differently? >> I mean most meetings can be replaced by an email, right? But maybe if there's a way of having the AI just everyone submit what they're working on or or don't even submit, just read what their chat transcript. That is so not fun. Like if if reading what someone is working on. Oh, they're working on this. >> No, but the reading wouldn't be to read. No human would read it. It would just be for the AI to decide which of project is most interesting to do a deep dive on. >> Okay, I see that point. Um I I think it could be used for that. But when you only have seven people in the room, you could literally just do >> that helps a lot. Yeah, if it's just seven people >> and you know the Amazon rule of like two pizzas, no no team is bigger than two pizzas. So have you heard about this? No. >> So the idea, this is Bezos's idea that any project that gets done, you should be able to feed that team with two pizzas. So that is about six, seven people and that's you don't have a lot of distractions there and you have a lot of responsibility. So once the team gets to 1215, oh I don't know what is that person working on. >> Mhm. I have a million questions but I I'll just start with number one. >> Yeah, I'll start with the most important one. >> I totally agree you keep the core team small and focused, you know. Um the the I I think a question number zero is how do you identify that small team? >> Oh, I love that question. >> Is is it self- select? Is it because I I I dealt with both. I I I dealt with you know I talked to the CEO then the CEO say these five people will work with you. >> Yeah. >> Sometimes it backfires sometimes it work great. >> Uh yes. >> Okay. >> Yes. I agree. So if I were selecting completely from scratch an AI center of excellence that huddle team I would say who is who are the few people let's say five to seven people who are working on the biggest problems using AI or AI native and biggest problems it has to be revenue or cost driven it has to be something that affects the company not just how a process in one department improves so if I were doing it from scratch but I've run into a lot of the same things you're saying where I thought this might be a very strong person but it ends up it was just peripheral it wasn't Not because the person wasn't strong, but for the things we're trying to do, the goal wasn't big enough or the way they were going about it wasn't as AI native. But we don't know that when we're in learning mode. So what would you do if you were setting it up from scratch? >> It's very interesting because you know what I you you keep saying there has to be attached to the core business. I was taking the totally opposite approach. >> Oh, I want to hear this. Okay. >> I was working with people has nothing to do with the core business. For example, I would I would mostly start with the supporting function, the HR, the finance. Not saying they're less important, but they're not touching the core business value as much as you know what you described. The the I can see your point. I I'll come back to that. But my point being if we fail, we fail softer, right? when we're not tapping into the core business. So if this thing doesn't work um then they don't lose too much and the CEO actually >> Can I reply on that one? >> Yeah, >> you don't want to fail on your core business. Agreed. Completely >> you can pilot things. >> I love that. So but I guess that becomes a question. You know, it's easier to convince the CEO, we're experimenting on your non-core businesses rather than we're going to tap into your core business and there is a 20% chance of failure. >> Jay, I love this back and forth. I would say convince your CEO on the core business. I would love to learn how. I would love to know because a lot of times when you start talking about their core business and because of you know all the you know fastpacing and uncertainty with AI they tend to say uh I don't think so. So, but when you start with I'm going to help you with your a HR function where you know we can you you can you can actually literally save money there too and then you can relocate the money into your core business and do something else and you'll see the success they tend to say yes easily. >> Can I ask you both um you haven't mentioned it but I'm wondering if this is on the back of your mind. It's certainly in the back mind hearing you. Uh if you tell somebody I'm going to tweak your HR or marketing or finance, they might think with AI that you're going to lay people off and like fine, I don't mind losing those people, but I don't want to lose the core business people. These are my family. I've grown up with them, right? They've contributed to the growth of the the company. They're they're here for the long run. Does that enter into what you're talking about or no? >> For me, it doesn't. In both cases, uh, when I've looked at it from an a people lens, whether it's a support function or the core business, the first question is is not where is not, let me emphasize not is is not where are we going to put in AI and take people out because so many people everybody can be upleveled. So in both cases, it's who are the people that that are doing the right processes? How can we make those right processes even better? But I know what you're saying. So let me not discount what you're saying. Yes, some people would be t in as in any course of business. There are some companies that annually exit 5% of the company just to be to make sure they have the most excellent players. Yes, in all over the world that is a concern. Are people going to be taken out? But that's not how I would position the project whether it's within these businesses or a pilot within the core business because you want the b that business to elevate and then you can make decisions. I I I would love to talk about that, but I think that's for a later conversation. I want to address what what you said. I I I promise, you know, I I come from your viewpoint. Start with the core business. I can totally see the value in that because that's how I manage my own projects. My own projects always follow a fail fast approach. So, because you know, I'm the CEO of myself. I I always say yes, right, for for new project ideas as long as they're valuable. So I would start with the most critical most loadbearing part of that project. I'll give it a try. If it doesn't pan out, then I can kill the project. So in that sense, I can totally see the value of your suggestion start with the core business. But I would really love to learn how do you um convince them, you know, it's not as risky as that. But maybe that's that's a that's a that's another conversation as well. But another thing I want to say I think it to to your earlier point is you know you want actually the commitment you start actually with the commitment from the executives before you can actually do anything. I I totally agree. I think I learned from you know some very experienced consulting people is within the business you want to carve out a piece called no going back zone. Uhhuh. That's interesting, >> right? So no going back zone means no matter how it turned out, this this is something you have to do. So if you find this no going back zone collapse with correlate with AI, meaning AI will help with something in that zone, then you start with that. >> Uh can you give an example of a no-go going back zone? So um for example um like you said you know saving cost you have to do that doesn't matter if you use AI don't use AI you lay off people don't lay off people you have to save cost and same thing goes with increase revenue and of course you know that's very abstract every business will have different ways to to save cost let's say you have this way to save cost by you know let's let's take F's idea we're going to replace 80% of meeting with with emails and we decided that will save cost. So that's your no going back zone because without without AI you do it. >> That's great. Yeah. >> Yeah. So that that's been the mentality I'm trying to use. But um with AI without AI or or maybe I should say before AI it is easier to create your no going back zone because most of the tech are proven you know they they established you easier to convince people this would work but with AI they will say okay I don't know I don't know what that is I don't know and even particularly with more tech savvy or aware people they say You said this will work this week. I don't know if it's still going to work next week. >> So, I think it's a lot of I as as Phil said, I think you're the perfect person to talk about this because you're the chief people office. I think this is 80% a people problem if if not more. >> I I'll tell you why I believe it is largely a people problem. I wouldn't even call it problem a people situation. >> Yeah. If the people team is not helping move AI initiatives, then these initiatives are not going to get out to everyone. So we it's a little bit like internal comms. And I love internal comms because internal comms is not sending an email. It's not sending a Slack message. It's not having one all hands. It is figuring out a way when everybody on every team, their manager knows what they're going to say to them at the meeting. They're going to get feedback from that meeting and they're going to circle that back and say what's going on. It's so I think sometimes comms gets a a bad rap in that oh it's just it's a one-way information but what if it's two-way information so I agree with you people if people are making change that's effectively a people department situation and for AI to stick I don't think it's an IT issue I don't think it's a CEO issue I think it's the kind of thing that you will want maybe you want a small more core part of your company using AI first but eventually in right now you will want your whole company using it >> and And actually one of my um I would say most or fastest moving client is actually come to think about it is it fits a lot of things you just said. These people reach out to me because they are an analytics function in a hedge fund. Arguably you can say either way they are the core business they are not the core business you know you can you can argue either way but they don't think they are the core business they think that actual financial research part is the core business in in a hedge fun of course you know I know this much about hedge fund so I don't know I I you can argue either way but they feel like you know what what they literally told me is we don't want to lead the charge in the a in this AI wave in my company, but we don't want to be the last because we're techsav savvy people. So, we have to do something and they literally started the next day doing something because I think that's motivation. So, I don't know, you know, core function or not core function, as long as they have a a set objective and they actually have motivation to reach that objective, I think that will get things moving. I would love to learn, you know, what do you guys think about this? I want to know about successes and failures and how do you know if you're early or late to what you're talking about? Like if you're from all the companies that you've seen, whether it's ones that you've worked with or worked at or or led or have heard about, what what are the factors that has anyone regretted trying to go AI native? Has anyone regretted starting late? How do you know when it's time? And why why like if it's such a great thing, right? Why isn't everybody already AI native? >> That's such a great question. Why isn't everybody already AI native? Why am not am like at an individual level? Why am I not completely AI native in the work that I have as the >> founder of Silicon Valley change in the fractional chief people officer work that I do. I think it comes back to this tension that everybody has two jobs. So let let me tell you both what the tension is, but let's brainstorm how can we get around this. I have a job as a chief people officer. You have a job. You have a job. You have to do your job and at the same time you have to be in charge of the transformation of your job to an AI native job. That's two jobs. So either you've delegated almost everything out of your entire job where you're the expert and the subject matter expert or you're still doing your full job and you're figuring out how to do a transformation. That's all doable and everybody is doing it. Every serious business leader is doing it. But because of that, you you can't jump from here to learning Italian completely. You just can't because you still have to be speaking English the whole day, for example. I want people to. So, let's figure out the other way. If you let's let's take an assumption that everybody has two jobs. Their job that they're amazing at, that they're senior vice president at, and they have to transform their entire organization to more AI native. Some people are going all the way, but to more AI. What do they do? >> Well, didn't you answer this? Pilot, right? Try parts of what your workflow, right? automate it and then just do more and more >> pilot try parts remove parts I think part of it has to be removing parts so one of the my favorite things that I've seen in both work that I've done but what my fellow chief people officers have shown shared with me is remove entire processes like don't even do this process uh and it starts it starts from a great place it starts from ideulating like let's say I'm in the people function and I want to know what's the best candidate experience when they first see awareness when they first see us when they first come into the company when they have interviews. What if I change that completely? And if by changing that I take out a whole entire processes that I think is also great. So you can pilot things, but what if you could remove just a bunch of spaghetti that's been on the walls for a long time. >> That sounds awesome. I think um just as a shoutback to our loyal listeners, the very first podcast episode, Jade said something that's been ringing in my ears and my head ever since. And it's a lot of what you're saying, which is AI isn't just about replacing people or doing what you're doing a little bit faster. It's removing entire processes, elevating the way you think about things. It's a it's an opportunity to reimagine everything, which is reasonable and that's wonderful. That's the world we all want to live in. Um, what what has been preventing people from from getting there? It's not hard to pilot. It's not hard to figure out what processes you don't need, right? Every single individual step is doable. So, why are we not done yet? Or of the people, what have been the biggest success stories that you've heard about or or or architected and orchestrated? um even if it's only for a particular department or something. What are some metrics that we can see that this has worked and it will be a shining city on the hill for everyone else to follow. >> Let's go metrics first and then uh success examples. So metrics, I love metrics. Efficiency is the simplest metric, but I don't think it's the best. The best is going to be a revenue and cost. So a profit metric. So any metric that moves the dial on cost and on revenue is it it's the best for the company. But you may not start there. You may start there with efficiency. What I've seen and colleagues of mine in in that run HR departments is cutting process time from 5 hours to 10 minutes regularly. And if you do that for hundreds of people within your organization, let's say everybody has to do the same kind of process once a week. If you can cut that and you can that's the easiest the most natural spot. So that's that's great, right? You have a first metric. They're saving time. What are they using that time towards? Okay. Now, that time can be used towards more strategic things on cost and on revenue. So, metrics, let's start with efficiency and then do things that actually, oh, if we do this, we're going to save $200,000. If we do this, we're going to bring in an additional million. Just anything where you have that goal, but can also work towards it. But comments on metrics? >> No, that that sounds fine. The only thing it seems to potentially be missing, but it's probably reasonable at this time is investment, right? There might be something in the future. It doesn't bring revenue today or doesn't reduce cost today, but we think we're building towards something. But I agree with you. There's so many lowhanging fruits. Let's just do the quick things now. That's reasonable. I >> I see what you're saying about investment and that will probably be as the organization changes as well. But how are we what's the organization of the future? There's this beautiful thing. I'll send it to you. Maybe we can put it into the notes. There's a beautiful study uh out of a Stanford researcher when I went to a conference of how organizational design may be completely different. We're so used to looking at org charts like here's the flat org chart. What if it's a three-dimensional org thing that you can visualize when a whole team comes in? It's beautiful and you can see it. when a team is has less power in the organization, you can see it dwindling down. But like just but that's valuable, right? For for someone who's making decisions, we don't even think about it that way. That's what I'm trying to say. Like I saw >> what what is that third dimension? So the top dimension is just hierarchy then cross functions. What's the third? >> I will show you a video of it because the video is the clearest thing. It's it's not it's not just how long have they been here or what are they working on. It's sort of where are they coming in and where do they have investment and what are the outcomes that they're doing. >> So it's it's kind of a combination of all that but we don't even think about that. We think as org chart is or chart P&L is P&L. >> So it's more of a graph like input. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Uh so that's just on the orchart but you oh we were going towards success. Yeah. I'll tell you an example that I got envious of. I went to a chief people officer conference about a hundred of us. Great conference. I recommend it. If you love that kind of stuff reach out to me. I love the company that organized it. And I'm speaking to this one person. and she's a chief people officer at a high-tech company in San Francisco. And she says, "Here's what I'm doing." And this was before I started doing it. So, I got I really did. I'm like, "Oh my gosh." She says, "I have a" which is so common now. I have a chief of staff, an AI chief of staff. I have AI processes. So, when I come to work, I push one button and I drink my coffee. My chief of staff tells, "But now everybody does this." But when I heard about it months ago, I'm like, "Oh, wait. I haven't created mine yet." So what I'm saying is success is probably seeing what hasn't been developed into open AAI yet hasn't developed into chat sorry into cloud because these are now things that are just features but it's what's that thing that's ahead so you can be there just a little bit ahead of the time. So for me it was really impressive that she would do this then she would look at this one process another she's basically managing everything uh from a bird's eye view and this is months before the rest of us were doing it. Well, let me I I feel the same way. I I I want to be on the um cutting edge. On the other hand, building it from scratch versus waiting for Claude to roll out a better version in two months. I'm not sure. I mean, how much time is it going to take me? You know, why not just wait? >> No, that that's reasonable. If you're if you're focused on other things that are bringing you value, that's that's reasonable as well. >> Yeah. >> So, okay. So, success. I I like the success. the the biggest thing to me of what you said, the success that sounds um enticing and inspiring is uh removing all of the drudgery, right? And being able to um have the most important problems surfaced to you so that your brain is working on the things that you enjoy working on that you add the most value to the company to. That that sounds to me like I'm not sure there's a a KPI for it, but um internally as a human being, then you know you're fulfilled. you have purpose, you have meaning, you're contributing, you're you're you're in the right place for your humanity, right? You're not doing the stuff that you don't like that's been outsourced. Um, it doesn't necessarily translate into revenue. It might translate into more risk even, right? If the AI is doing processes that you used to meticulously check over every with a fine tooth comb, it does still make mistakes and it might take a week or two weeks before you find a particular mistake. By that point, you feel like it's been going well. So it could be that AI native is at times an illusion whereas fragile in a way that you might not be able to tell later but I still think it would be a success because you're able to elevate you know to use Jay's term of the kinds of questions you're thinking about. >> I think about this a lot what you're talking about. I don't know what the resolution is but one of the most common complaints from the executive coaching practice that I've heard and a valid complaint is I don't have enough creative time. I don't have enough creative space. Yes, I now have lots of agents, dozens of agents. They're working for me. I'm overseeing them. I don't have that space to figure out how the business can be exponentially better, not just procedurally and incrementally better. >> Is that something that is a goal of AI native or that's orthogonal to it? >> Oh, interesting. I don't know that that is a goal. The goal of AI native is for the company to be able to transform when it needs to. But I don't think a goal of AI native is for people to have more creativity time. I think that would be a very interesting correlary. But I I've never >> So what is the role of people then in an AI native firm? >> It still to make the dis first of all it's to drive the processes. So you >> why >> why doesn't AI just drive the processes? >> It can drive a lot of the processes >> going back to your four points. Yeah. >> Right. That's that's humans today, right? At every level deciding to go AI native. Why is it humans? If you're going AI native in a sense, it feels like you only need to make one decision, >> right? There should be a a clawed AI native thing and it reads all your stuff. You flip it and it does it for you. >> Well, it's really interesting. >> Both of us drive a Tesla, right? So, I think it's your decision when you engage FSD, do you let go the steering wheel? you as the human >> make that decision, right? I thought you were gonna say we decide where to go. >> We don't I think engaging FSD is the going AI direction. >> Okay. >> So, but you still hold the steering wheel because you still drive >> you still drive. >> Do you want to hold the steer? I don't want to hold the steering wheel. >> That's something I think we should talk about next. Okay. >> But right now, I I I think I want to comment on what you both just said. One is you know Phil what you said you know why don't we just wait it now becomes a rule of operation for me that if someone come to me and say build me this and my judgment co is this is a rapper of JBT or cloud which means you know they can build this in two months I'm not going to build it I'm I'm I'm going to directly say no even if it's a valuable thing right now because I think it to 's point I think it's a waste of my time right and and and and sa to your point I I guess there is not only an investment you know cost there is also a lag the lag from somebody who build AI product every day the lag mainly um exist because I don't know what's the most efficient way to transfer what's in here into AI that's actually what I do every day right now my my academic research my practical projects is all about how to transfer human knowledge into AI because um that that's a talk I'm I've been preparing for for weeks now is I did everything the book said you know you should you know um plan first you should you know let AI interview you should create a scratch pad I did everything I actually have um over 15 rules in my harness that that guarantee the buy the book knowledge transfer but my agent is still not there. It's still not I guess we're going to talk about replacement and and co-pilot next but it's still I don't think it's ever going to replace me. And by its own admission every turn of conversation is say okay in this session the most valuable you know comments come from you not me meaning come from the human not AI so I I think you know we're hinting on this thing but we're not you know really touching on this and what promise our audience we're going to talk about this is the human role in the AI native business. Yeah, >> we we we use different terms for this. We use governance, we use oversight, we use co-pilot, we use fields, you know, all hands off kind of thing. So, what's your thought on that? How how how would a human be in the AI native environment >> that when I look at what's popular and what people really want to know about how AI is progressing, the only two things that I see in terms of where is a human involved and it's the same thing as governance or oversight. So, We create these great things, all of our individual agents and then our corporate agents, but how do we know that they're going to keep doing what we want them to do? I don't think we're yet at your world, which is a very interesting world. It's still intriguing for me. I don't think we're at the world of, hey, we're just going to let it all let it all do what it needs to do. So, it is the governance and the oversight. And it works in lots of different ways. It can work as a very routine cadence of at this point we or look at what we really need or it can work with a pivot that you want to do with the company. So if you're the executive team of the company and you want it to move in a different direction, how as part of that pivot are we going to change our processes including our AI processes >> and and and there's another term I think is also very important and I think that would be the key for you know the the the dream word of fulfills to happen is alignment is is the human alignment that I know the agent is doing is supposed to because he's doing it my way so I don't have to book. Well, if it's an AI native firm, we talk about alignment as if the AI needs to bend to human will. But shouldn't the humans be bending to the AI will for an AI native firm? >> Bill, you are so far ahead of the firms that I work with. >> I like it. I'm just I keep being intrigued, but it's it's not as practical with the executives, whether we're coaching or I'm inside as a on the people team. I I I I'm drawn to what you're saying, but I I don't know what to say. I don't want it to be an obstacle. I want people to get to your vision, but I also want people to still guide the AI where where the AI is still developing. >> There there was one example of a company, I can't remember the name, that was completely AI native. It's it's an agent and or sequence of agents and it needed a human CEO. So, it put out a um a hiring ad. We need a human CEO. You won't really do much. It'll be kind of a figurehead. But if you can persuade the AI to pivot, to change, to do whatever, then that's fine. So the alignment doesn't have to just be we got to pound on the AI. >> What happened? >> I think they hired someone. I don't know, >> college kid. >> Greatest job of all time, right? You're just persuading AI to do what you think is best. Sometimes you're wrong. >> Yes. >> Isn't that great? >> Yeah. >> But isn't that some kind of alignment? >> Yeah. So the Exactly. So it's both ways alignment. You have to align to the AI, not just the AI to humans. >> It is. So I guess the human role is to me you right now I I'm I'm far from AI native you know even in my own little isolated environment I'm far from my AI native but my way is it is a co c co-pilot of me sometimes it it takes over for me sometimes I take over and it work with me it's it complement me it doesn't replace me >> I think that's really cool because I think a lot of people are not using it as a co-pilot as colleague, they're using it as a do my mundane tasks, which is also very effective and very important. But I think you're elevating the AI's work when you work with it >> because I'm not working with one AI, I'm working with 50, 100 of them, an army of them. So they I can dedicate a function of those. So from overseeing to co-pilot and hopefully to feels, you know, all hands off kind of a dreamland. Well, let me ask you guys one last question. uh from your different perspectives. I think your perspective, correct me if I'm wrong, more on the top of your four features, right? Closer to the strategy level if you're coaching executives or from the people level like kind of top down and yours is probably more in the huddle, middle out, right? What advice would you give to companies or the people within the companies who assuming they want to go AI native? Um what advice would you give them? Pitfalls that you've observed uh what what should they wait? Should they not wait? Any advice? I would say learning mindset and uh pivot pilot/experiment. So what I mean by learning mindset and this is something in psychology there are completely two different things to look at in your business outcomes when you're in a learning mode and when you're in output mode. So most of the businesses that we know that have great products they are absolutely in output mode. I mean as a mindset. So it's something it's you're switching to this is investigative, this is experimental, what do I do? which means you're going to have different KPIs. You are not you're going to have did I figure this out? Did it work? As opposed to did I reach this cost savings. So that's one thing. And then when I say pilot or experiment, I actually mean something small. So try it as opposed to just have AI plan it. Have AI try have AI help you as a business try it. So both think about it in differently. Don't don't don't say no to the AI because it didn't give you this X of costsaving early on because you just don't know. You're experimenting. And also try it in small different ways. >> Makes sense. Well, that's ex excellent. I I think there's only one thing I can add. That's from my own experience is open-mindedness. I one example I give you is I let AI ask me questions before it can response anything back to me. Right? I know everybody use that and and everybody say you know that's that's a great pattern to use but it actually changed me I think as well as the AI it start you know I start thinking about why didn't I ask this questions myself >> what am I putting off here why am I why didn't I start thinking about these things so it started think now when I start a topic I I would directly think oh AI would ask me question one two and three I better have answers for those. So now I think that I my planning skill got got better. So it's it is indeed a two-way street. That's awesome. >> Yeah. Thank you SA for joining us. Uh the uh founder of Silicon Valley change uh chief people officer in AI uh Harvard UPEN Stanford uh greatest older sister of all time. >> Yes, that's the only one that matters. >> So thank you for joining us. Uh this was a great episode. But um we learned a lot about AI native uh in the in the macro right of the big firms and how to translate them and how to move them and what's happening in the world today. Uh we'll have another episode we'll see you in a few seconds but it might be a little time for you guys at home. We'll have another episode talking to SA about uh micro uh AI native. How do you do it as an as a soloreneur as an entrepreneur as a new business today where you don't have all of the legacy stuff that is important? How do you move things into the future? What if you're starting at the future? Thanks for joining us.