Maximos Lih - When One Person Does the Work of Seven

When One Person Does the Work of Seven: How AI Is Rewriting Work

What happens when one employee can suddenly accomplish the work that previously required an entire team?

In this episode of the People/AI Strategy Forum, Sam Reeve speaks with Maximos Lih about how artificial intelligence is fundamentally changing productivity, workforce expectations, and the future of organizational design.

As AI expands what individuals can accomplish, leaders are being challenged to rethink workforce planning, talent strategies, management practices, and how organizations create value.

The conversation explores both the opportunities and risks of AI-driven productivity, including how organizations can embrace innovation without losing sight of the human side of work.

In this episode, we discuss:

• How AI is increasing individual productivity
 • The implications of AI for workforce planning
 • Why traditional organizational structures are evolving
 • The changing role of leadership in AI-enabled organizations
 • How businesses can responsibly adopt AI
 • The future relationship between people and technology
 • Skills employees need to remain valuable in an AI-driven workplace
 • Balancing efficiency, performance, and employee well-being

Key Takeaway

AI is not replacing work.

It is redefining work.

Organizations that understand how to combine human strengths with artificial intelligence will be better positioned to adapt, compete, and grow in the years ahead.

Guest:
 Maximos Lih

Learn more about CompTeam:
 https://compteam.net/

Watch the full video episode:
 https://www.youtube.com/@PeopleStrategyForumPoweredByCompTeam

#PeopleStrategy #FutureOfWork #ArtificialIntelligence #AILeadership #WorkforceStrategy

If you enjoyed this episode, follow the People/AI Strategy Forum on your preferred podcast platform and join the conversation! 

About the People/AI Strategy Forum
The People/AI Strategy Forum explores how leaders navigate the intersection of people strategy, leadership, and artificial intelligence. Hosted by Sam Reeve, Founder & CEO of CompTeam, the Forum features conversations with executives, practitioners, and experts shaping the future of work.

Learn more about CompTeam and the People/AI Strategy Forum at compteam.net.

Transcript

Sam 0:00

What if AI is not just change... about changing jobs, but, but breaking the entire logic be- behind how we hire people, retain people, and pay people, specifically around performance? Welcome to the People Strategy Forum. I'm Sam Reid, your host and CEO of CompTeam, where we help organizations design people-centered compensation and career systems that attract, retain, and motivate. Today's topic is one that leaders are not talking about nearly enough. When an employee is equipped with the right AI tools, they can produce the output of three, five, or even seven traditional contributors. So what happens when we are, are not paying them effectively? Uh, does that impact, uh, our retention? And what do we do with workforce design? These people can easily be poached, so we need to rethink on how we approach this problem. Our guest today is Maximos Li, founder and CEO of Emboldened. Maximos brings, uh, an unusual blend of startup and talent strategy, org design, executive coaching, and operator-level excellence that, uh, uh, some... I- is difficult for others to replicate. So we're really glad to have him here today. He's advised hundreds of growth stage companies, helped them build hiring systems, people systems, and work at the intersection of talent, execution, and business transformation. So today we're digging into the questions: How should leaders reward and motivate people when their value comes more from leverage than hourly labor? So let's welcome Maximos. Hello, everyone. Really glad to be here. Yes. Thanks, Maximos, for joining us once again. I know we've talked before on, on the, uh, uh, People Strategy Forum. I'd love to, uh, uh, just dive into your expertise around what you're working with right now with AI-empowered startups. And so, uh, I, I know you're working with, uh, plenty of those, those types of firms and, and, uh, leaders. R- is that right? Yeah. I would say that where I've had incredible privilege is I spent almost a decade working on the team at Google Ventures, and my job really, you know, even though it transformed a lot during the time that I was there, um, really was about how do we maximize the rewards and returns of the checks that Google has written out to startups they, they find interesting. And they could be, you know, anything from health tech to Blue Bottle Coffee to Nest and Uber and Slack and Stripe. And so what I often have the privilege to do is get a frontline seat on companies that are trying new things that are really taking less is more to an extreme. Mm. And what I think is so fascinating about the topic that we're discussing today is that- My mandate in 2017 when Google and Google Ventures were mainly interested in series HFs, right? Where $10 million raises, 30 to $50 million valuations, was if I could help you get from 50 to 150 without growing, build, breaking your quality bar, and 150 to 500 without breaking your culture, you will generate returns. That was almost always the intersection of people and revenue. And today, since I started my own practice three years ago, we're seeing startups push automation, tooling, less is more, to an even higher degree, not because they're outsourcing, which would've been a cost-benefit leverage, you know, in the past, but really saying, "We have agent-based tools now where I could be a team of 12 people with an ARR of $30 million, and I don't plan to be more than 150 people at the time of exit," right? And so it's like, oh, is it actually possible to say 50, 100 person team can get to $100 million ARR and either get a very, right, profitable M&A or actually a, a worthwhile IPO in these specific spaces? And so yeah, right, if you're working at the intersection of people and revenue, and what you're gunning for is $100 million- 100 million ARR with less than 150 people, everything about the way that you hire, right, evaluate, and retain talent will become unrecognizable from the old days when it was like, how do I get you from 150 to 500 without breaking your culture, and maybe 5,000 if you're thinking international. That's really interesting. So let's dive into that a bit more in, in practical terms. So when you have a, a new startup that comes your way and they're trying to, um, um, uh, be a force multiplier and with their people and they're... So what's their, their hiring strategy? Are they... They're looking for... I imagine they're looking for people that are re- really AI-enabled. And, uh, how do they set up their teams and, and ensure that they can be that force multiplier? The two biggest differences that I've seen really are happening at the Series A and Series B stages. And for non-startups, I would say, you know, the first 12 to 20 people that you're hiring when you're just trying to build something that doesn't break and not run out of money Yeah are generally your friends, right? They're known quantities of people that you've worked with, who have a personal relationship with you, who are willing to take the pay cut on a mission, and also just the fun of building something in a garage with your friends. Mm-hmm. You're not really thinking about them as force multipliers at that point. You're just trying to think about them as, you know, people that want to do this adventure when you don't know what you're doing as a founder, right? But it's really past that 12-person team, right, when you're like, "All right. Maybe I think we've got a product we wanna try to scale," or, "Maybe I've raised a little bit of money and now I've run out of friends and the expertise that I need to bring in," that you start to actually have that question of, right, what is the first person in this function? What is that role supposed to be like, and how do I actually get them to be better at what I was doing, right- Yeah in the job? And, and sometimes I had a great mentor named Joe Kraus who used to say, uh, "If you're a founder that's not willing to fire yourself from a job, it's because you haven't seen true greatness in that area." Mm-hmm. Um, and so I think that that's kinda where a lot of my work, as well as where founders are really getting the most leverage. So two things that are happening. One, um, and we've spoken about this before. I think it is a gross mistake for founders to be hiring for experience in an AI and agentic world. And it may surprise some of your readers to know that as of this very moment, the internet as we know it is 53 to 57% agents and not humans, which means that the internet is now agentic. So the way that you're trying to get leverage on the internet- is not the human first kind of behavioral practices that you think you were working with before. So the experiences of being able to optimize for that may not be very valuable as the agentic internet is growing to 60, 70, 80% agents over humans. Um, yeah. So, so one thing, let me just interrupt you really quick there. Yeah, of course. I mean, when Can you, can you, uh, explain that when you say that the internet is more than half agentic now, what do you mean by that? The traffic and the causes of traffic are being done through agents more than they're being done by humans, and you could track that even through, um, you know, uh, trading on, you know, the, the stock exchange, right? How much of those transactions are happening because of agents that had rules programmed into them to auto balance for a certain, you know, percentage yield or protect against a certain percentage of loss as opposed to a human that is watching Twitter, looking at analysis, reading the graphs and the candlesticks and then, like, making the plays in real time. More than 50% of them are now agents that are just automated. So I would expect that if that, uh, if that's a, uh, um, the, the reality that we should s- have seen that the internet has expanded by, uh, two times since the, uh, the fur- in the past five years. Is that, is that true? Is that true? I would say that that is fair, but don't underestimate that that's the number of people who are now not actively watching X or, uh, and candlesticks, and now have the ability to go do something else with the time once they put those automation rules in place. Nice. Um, and so I think that that's also where we're when we talk about leverage and headcount and team size, yeah, it's 'cause if I write the right rules, my SEO, right, my manager is just going online looking at how I optimize for the Google or Facebook ads that I purchased, and I am not writing content. I have, you know, a content calendar that Claude is actually churning out, right? And then I have, you know, somebody who's looking at videos and cutting them up into things that I can drip onto LinkedIn. And so then what I'm actually doing is deciding the right import- like, what is the prioritization of things. But I have agents that are doing three or four jobs that used to be occupied by, you know, uh, three people between three to seven years of experience. Right. So let's go back to the, the, the s- the small startup example that you were talking about where we've, we've, uh, got our, our core team of, of humans that are work really well together, and now we're trying to, uh, be a force multiplier. So what's the next step that an organization can do to, uh, really get that next level of productivity? Yeah. So, so I would say the three mistakes that most companies are making right now is, one, they over-index on experience, and they're not hiring new grads. Um, and that's, that's a mistake on a variety of different levels. Um, but predominantly just because the experience that you think you would get from those people actually may not create any leverage in an agentic-first ecosystem. Two, they are, uh, still keeping people in very specific domains. So you think, "Oh, if I'm hiring an engineer, I should hire engineers to do engineer things and measure them by lines of code," correct? That is untrue. Actually, the engineers of the future are always going to be product or UX engineer hybrids because the one thing that agents cannot do, will likely not do for the next 10 years, is actually be the purchaser. So we are not at a place where we're willing to trust an agent with my credit card information and just say like, "Buy the thing," right? But they will do everything short of that. Um, and, uh, so then the purchasing behavior will still have a human-based component to it. That means that it's gotta be a UX-based or a product-based mindset, and more and more lines of code will be pushed to Claude, to AI tokens. And so my most highly leveraged engineers are the people who understand how something works to get a buying behavior, not the person who writes the most lines of code. And then the third thing is, um, that they underestimate the, like the value of systems thinking. And I'll give you a very specific, you know, concrete example of that. I have a company, they are $30 million ARR. It's a 12% engineering team, 20% total team And their engineering team basically has a system where three senior architects are creating a cloud-based framework of how features can get suggested and then translated into prototypes and code. And that, uh, layer has a supervisor layer that says, "And every time that you actually push out a feature, I want you to do your own unit tests that are totally automated, because I want a zero technical debt system. And then I want you to be able to create agents that go to watch my junior new grad engineers when they're trying to push things out, and run unit tests based off of them. Anything that comes out with a bug, rank that bug on a scale from one to 10. If it's a level four or below level of difficulty, fix it yourself and run it until it's done." Mm-hmm. "If it's between a level four and seven, um, optimize the algorithm that you're doing the rankings for so that you could be able to fix those and push them down to level three and below." Mm-hmm. "And anything that is level eight and above, send it to a senior engineer, because I want a human review before it gets pushed to our code base." Mm-hmm. Right. So they're able to hire 20 new grads within a system with very little understanding of judgment, but yet a massive amount of productivity. And it's these architects that are not thinking in terms of how do I build something for my own agent, but how do I contribute to a platform-based agent with a supervisor laser, a, a supervisor layer of agent that then makes everybody in the company more productive. Okay, so you just... L- Let me just make sure that this is clear and understood. You, you just talked about a structure where, where, uh, AI is pushing, um, uh, uh, the, the most complex pieces to the senior engineers. And, uh, right now we're hearing that there is a lack of hiring in organizations, um, uh, for those e- entry-level jobs out of, out of college and so forth. So are you saying that those entry-level jobs that in this new future organization you're talking about are, are those that are empowered with agentic knowledge, that they are the agentic m- m- managers? And, and then we have the senior engineers that are just more specialists? I would say that the movement needs to be away from let's give everybody a $500 ChatGPT budget to make their own agents. One, that doesn't actually make the company work better or faster because it's not thinking in terms of workflows. Two, you could have agents that end up competing against each other. And three, those agents don't belong truly to the company. They can be easily replicated somewhere else. So, um, that system is actually not incentivized to make the company better or more productive. So what you need is layers of agents, right? That actually act as governing forces over entire workflows. And in that kind of a system, right, where an engineer is understanding the right features and use cases, but not necessarily running their own unit tests, um, a mid-level engineer with seven to 15 years of experience is actually going to be, uh, ill-equipped to do that system. They will be too tempted to fix the mistakes and get frustrated at the hallucinations that the AI creates, rather than saying, "How do I actually teach and train the AI to write better, you know, rules so that I never have to run a unit test again?" Right. Gotcha. All right, so, so how are the organizations of the future setting up their, their teams and, and the, uh, with the agentic teams? Are, are you, are the, um, are the agents used across the organization, uh, uh, by multiple people? Or are, are you seeing that there's agentic managers that are responsible for a group of, of agents? I would say that what you have are agent trainers who are responsible for creating a layer of institutional knowledge that belongs to the at the company level, is not owned by any single individual. So it's not your knowledge, it's any time you have domain expertise or knowledge, you're feeding and training the company agent. Um, and then you have, you know, quality control agents that sit on top of that. So imagine if I were to say, "Hey, I want you to write me a summary analysis of how crypto is working in 2026, and what you think it might look like in 2027." All of us know that it's just a matter of time before I hit hallucinations based off of that. So then I need to say, right, maybe one of three things, right? I need to write instructions that say you value accuracy and precision above being helpful. Do not say with confidence anything that is not 100% foolproof. And then I might say, I want you to create an agent that acts as a fact-checker, so that before you actually show me the results, the fact-checker agent will review what my analyst agent is writing. And then I have a third one that says, now imagine you are the board of my, you know, financial services LLC, because I'm going to be using these to write investment decisions. Now the board will look at what the analyst and the fact-checker have produced and tell me if you think this is a risk mitigated, right, a high-quality investment that aligns with my investing philosophy and my risk, uh, profile Right. So now you have three layers of agents, right? That might be coming from one person acting in very different roles, but the goal is to make a decision that I actually ultimately have, uh, hold responsibility for. Mm-hmm. Which is a very different thing, right, than just, um, let's say like outsourcing tools, right, to a agent at a single layer that's all flat and horizontal. It's actually creating teams underneath agent system. Okay, so one thing I just wanna make clear is in, in your structure, are you saying that the... Is there's... You mentioned that there's a, a number of human trainers that are training agents. So are we... Is there, um, is there multiple trainers training the same agent? Or is- were you looking at one human trainer training a series of a layer of agent? The layer, the... So the layer agent is an agent. So I think that- that's the thing that needs to be really, like it's still one agent. It's just a layer, an agent that is being trained to, with a goal to improve a company- Yeah as opposed to improve, like me, individual. Um, that has to be trained by multiple people- Okay for your company to retain its IP, and it has to be trained by people that you consider to have judgment. So it's not a democratized training, which is also the fallacy of let's... Even if you gave everybody $500 in ChatGPT to train one company agent, I don't want low performers training with the same weight as high performers. I want the high performers to set the rules because they're going to have better quality protection, and also a better alignment to the aspirations of where I want my business to go. Okay. Let's move to performance and how to measure performance of these, these, uh, future human employees. So, um, when somebody is contributing to training the agent, how do we know if somebody's more effective in that agent training than others? Um, I think that a lot of that is still actually being designed in real time. But a quantitative metric that I could use is for headcount planning. So if what you have produced actually does cr- like, allows me to not have to hire an additional head to do the work at the... of scaling whatever dimension you're scaling, speed, complexity, quantity, um, then that person is a high leverage person. And I, I tell companies sometimes we should actually seriously consider, and there are companies that are actually doing this right now, we just hold payroll static at 60%, and whatever we don't spend on new headcount goes back to existing headcount. So I could be paying a head- you know, an individual contributor the headcount of two, two and a half people because they created automations that could, like, save me money from ha- having to hire, onboard, and train a new person. And it is vastly important that I do actually retain that person at the value that they're creating for the company. Mm. Yeah, very interesting. Yeah, I imagine... You know, you know, of course, uh, being in compensation myself, there's a, uh, there's a lot of different aspects of how we can do this through, you know, gain sharing or, or, uh, incentives and so forth. But, but that's, uh, um, very interesting in how we're going to have to bridge this new gap of, of paying for, uh, focusing more on paying for output versus paying for the hours that we have a, a person doing a, you know, a, a job, uh, that we've been doing for millennia. Yeah. Right? In fact, I would say, and, and Meta demonstrated... I think Meta did some research that actually demonstrated the efficacy of this. But Meta did a research, um, around, like, work from home productivity, and from that they learned two things. Work from home disadvantages, uh, any junior hires, and so new grads, and anybody who is at what we call an L6 or above. Mm-hmm. Which is essentially anything that is more senior than a technical lead. And they learned that it was for two reasons. One, uh, the peripheral vision and osmosis that a new grad gets from being in the office and being able to eavesdrop, they're massively, uh, lowered when people are working from home. And then two, for those senior level technical lead hires, they are not doing the cross-functional collaboration and, uh, innovation that actually helps to move the company forward. And for that reason, they pulled both of those teams in, and then they started to measure the efficacy of being on site across individuals by saying, "My new grads should be, uh, writing a lot of lines of code because they're learning judgment, and they're learning, like, what good looks like. But my L6s should be almost not writing code at all." So I should be spending a lot more money on tokens, AI tokens, for my L6s and above because they should have so many cross-functional ideas that they are not dealing in the silos of actually doing that kind of, you know, um, mu- much, much more manual-based re- um, IC line coding. Um, and, and that's literally how they actually understand the health of those two job families, is token spend increasing as people get up to seniority, and is cross-functional collaboration increasing as they get closer to seniority? And then from there you can actually base then, okay, so then of the cross-functional projects that my good L6s are doing, which one of them are creating new lines of business? And that actually then, um, you know, f- feedback loops back into, uh, exponential compensation. Okay. Very interesting. So, so I, I wanna dig into what you just said a, a moment again, uh, a moment ago about, uh, the meta research. I'm a, I'm a bit skeptical about the, the, uh, uh, the h- hybrid or work from home, um, or being full on-site. So, uh... And I can understand why larger organizations may want to pull people back into the infrastructure. Uh, there's a lot of money, uh, spent on that infrastructure. But when you're working with, uh, small startups that, uh, you know, may have really dispersed teams and, uh, knowledge that comes from different parts of the world, I mean, it's, it seems to be a little bit more difficult to have somebody, y- you know, people come into the same space in that small startup. And so when we're looking at the, the nature of, uh, uh, on-site work versus fully remote work, what's your experience with, uh, your, the startups that you're working with? What do they typically do? S- small companies actually like being together. Yeah. Right? And it goes back to if, if I was one of 12 people and I wanted to build something with my friends, why would I not, you know, want the opportunity to go hang out with my friends when we're building something? So smaller teams tend to like it, and I do find that companies that want to move, uh, complexity without sacrificing speed, they do get better results from being on site. Now, there's a lot of job families and industries like enterprise SaaS or if you're doing government and public sector, where being in the same location actually doesn't make any sense because the market is dispersed. And so how do you prevent the insider-outsider issue, the home base versus, you know, a stepchild second, you know, office kind of syndrome? And in those kinds of cases, I think GitLab, which was a GV portfolio company, set a beautiful example, right? Um, where they said, you know, if everything is gonna be distributed then it has to be writing first. And I think that that's, that's really fair. You're not gonna call meetings because it was convenient. You create a document, and then when you have five to eight comments that don't get to alignment, you call a meeting with those people. The second thing that I think has been really useful, uh, that I encourage is don't mandate a two to three days in the office a week. You just end up missing each other or you end up going into the office and spending most of your time in conference rooms. If you're gonna be together, call for sprint weeks where everybody is in the same place for four to five days at a time. The cross-functional, you know, like magic that you want to happen happens when they're in the Uber ride back to the hotel. It doesn't actually happen in that conference room. It happens when they're going to the gym or at breakfast talking to each other. And you can only do that when you predictably know the person that you're talking to today is going to be here tomorrow. And that's a lot of scheduling if everybody is navigating their childcare. But if you do it for one week a quarter, and then everybody just go home and do your thing, but I know who you are so when you ping me or Slack me, I know to pay attention to your ask. That actually gets us a lot more productivity. I see. Now, um, l- let's jump back to the, uh, concept of o- org design, the future org design of organizations. Now it sounds like, uh, that, uh, you know, right now there's, you know Uh, uh, we look at some of the largest organizations out there, they have quite a hierarchy of different levels and, and so forth. And, uh, do you imagine that that might, uh, that stack might shrink in the future given, uh, the a- agentic, um, um, way forward? Most companies are already doing that actually, as I can tell. Um, you can see, uh, Anthropic OpenAI. They... If you look on LinkedIn, everybody is a member of technical staff, so they're not thinking in terms of layers or levels. Um, they're really thinking in terms of, are you an architect? Are you a trainer? Are you an implementer? Are you a business, uh, a customer-facing brand builder? And all of those things are gonna become much more... mu- much less, right, like striated because of experience. Companies that I know are also spending, uh, much less money on building and management layers because if I can have agents that are doing the work, then everybody is a manager, and I don't wanna pay the admin tax of having, you know, a specific front line manager. So I think both of those things are actually, um, contracting as we see it, and you can also see that reflected in revenues, um, for companies like Lattice, like Culture Amp, that are building management-based tools. L&D budgets are collapsing as well. Mm-hmm. So I can, I can imagine that, uh, more knowledge-based type organizations, technology organizations will probably be more that manager level. But then when we're looking at more into, uh, legacy manufacturing, it may be opposite. Is... You know, where we may have n- a more agentic type management or, or task management pushing down to, uh, uh, humans that might be doing the work or, or potentially hu- And, and those that are robotically, uh, en- enabled of course would be different, but I'm talking about those that require manual labor right now. That's an interesting question. So I... The, the majority of my clients are man- that have a manufacturing arm are doing hardware design and chip design. Mm-hmm. Um, I would say that probably 30, 40% of them overseas-based, and then the other 60% here in the United States. Uh, actually, I would say, yeah, plus another 10% may be in London, just because the London government is actively investing in that area. Uh, you would be surprised. I think so much of those types of manufacturing arms are getting pushed to foundries instead of companies wanting to take it on themselves. Mm-hmm. And then what used to be more test-based, quality-based, um, specializations, they're trying to either create or do design partnerships with companies that are specializing in AI-based design tools. So, um, I think it's still a very early stage of startup innovation right now, but the idea of if I can create a platform that is similar to Cadence but for, you know, quality and testing between when a spec gets locked and when manufacturing can begin, um, I can shorten that time from three or four months down to six weeks. The leverage that I get over my competitors as well as the costs that I would save or the headcount of those people who don't have anything else to do for the rest of the year, um, would be invaluable. And so then they are willing to pay, you know, 100K a year to try to invest in those. Hmm. Okay. Yes, so specialization, of course. So I wanna get back to, uh, what you mentioned as far as, uh, uh, the, uh, the flattening of organizations, and specifically in what, uh, career pathing looks like in, in the future. So when we, uh, of course, when we have a flattening or- of organizations, uh, the upward momentum of careers, uh, somewhat, uh, uh, changes its path or gets stalled. So what is your perception of what future career pathing will look like for professionals? I would not be surprised if most employer-employee relationships do not have a true W2 relationship in the future. Mm-hmm. Because, um... A- and I'm already starting to see this You-- anybody who is at the architect level could create an entire, uh, consulting business where I, let's say, I know what Salesforce is doing in terms of building CRMs and sales tools for different companies, but there's data privacy, and you're always tweaking it, you know, to a certain extent for your specific company's workflow. Well, with three months of time and really good cloud tools, I can go into any company that's 150 people or less and build you your custom instance of Salesforce. I can build you your custom instance of Greenhouse. You no longer need to buy into the larger platform subscriptions because I can build you one. Maybe that costs 50 or 80K for the three-month project. It's entirely done by three people with my platform of agents that are doing their own unit tests. So all I really need to do is figure out the product, right? Listening, workflow of that. And they can go from company to company and just keep doing that. I can be a specialist in internationalization and go from company to company, company, building out translation models for your, all your web properties. And, um, that no longer requires a W2 to get the level of compensation that you used to be at, nor does it really require you to have really good business sense because you're leveraging AI and sort of your n- your, your work product, right, is now this AI agent set that you manage and are able to take from company to company. Okay. So that's, that's, uh, um- So more contractors, more consultants in the future. Uh, does that, uh... Are those... Do you imagine those contractors or consultants being restricted, uh, in working with other competitors as part of the agreement, as a, as a w- you know, as a... Going to different o- organizations? Are you seeing that in agreements? Uh, I'm not actually. And, and I think that the people who are doing this early, probably what we'll have is either competency or project-based employment and less W2 title-based employment. It, uh, changes the way that companies do budgeting as well around skills and workforce. Um, and really the big lock-in for a count- country like the United States is then where does the healthcare, you know, benefits come from? But, uh, there's a lot of flexibility now in being able to say efficiency is gonna be a commodity if you- everybody is buying the same tools, and then it's actually your ability to use the tools almost in a rental model that's actually gonna get you, um, you know, your employment compensation. And some employers will really like that because it means that everything is a try before you buy review-based system, right? Not necessarily hiring based off of potential anymore, but, you know, like any other type of equipment or competency. "Hey, I'm a winery. I only need to bottle wines, you know, two weeks a year. It costs me, you know, 100K to do... To rent something, but it would've cost me $2 million to buy or build it myself. I'm gonna take it," right? Um, and, uh, you get an... And then they get to do it with, um, you know, sort of like an entity whose bug fixes are tried and true rather than the build it yourself where you're going through so many iterations of guess and check. Okay. I wanna, wanna jump on, uh, something else that you mentioned that's... It's a little off-topic, but you, you're saying that the, um, your, your, your example was, uh, SaaS companies. Yeah. And right now, of course, there's the, the, the big, uh, SaaS couplet that, that people are talking about, the destruction of SaaS companies. Is that, is that something that you see, uh, continuing on into the future? I guess, uh, I would be curious what you think and what do they mean by the destruction of SaaS companies? Well, you just mentioned as far as you made a couple examples as far as, uh, Salesforce, Greenhouse building internal, uh, systems that, uh, that replicate those systems that, that there's They can make it more customized internally. Right, right, right. Yeah. I, I would say that, that definitely what the, the incumbents should be concerned that the network effects of, you know, whatever If you're not getting a data layer for how I measure against my competitors in a way that's valuable to me, then being part of a huge conglomerate with an existing workflow may not actually be that useful, um, and certainly not worth the cost as I scale. Um, I think that there's going to be, right, like you're Any company that is not playing the I'm trying to use AI to be 150 people or less are still going to get so much standardization protection from using some of those larger systems, and maybe even just the internal data transfer and dashboarding would, would be useful for them to keep paying it. But for companies that are much more in the startup space, the build it yourself value will, will become much more clear over time. Great. Yeah. Hey, uh, thank you, Maximus. A- another great conversation. So as we kind of wrap up our, our discussion today, what are the top things that leaders should be thinking about when, uh, as, as the workforce becomes more agentic empowered? So I would say I'm doing a lot of headcount planning right now with my clients, and the question that I continue to ask is, are you sure that you're not hiring more people because that's the path of least resistance, but that that is truly the edge of your automation workflow? So I think that that's, uh, extremely important to ask. The second thing that I, I do is there's a identity change that needs to happen in the mind of the CEO. So I'd say I, I generally would tell people, when you're a series A company, if you're 50 people or less, you are the judgment, you are the taste, you are the voice of the culture, and you set the standard for excellence. That is a shorthand, right? Series A is just a shorthand for when the CEO is still doing some IC work and, you know, oftentimes acting as the first line manager. But in an agentic universe, you almost want to move out of that as quickly as possible and start asking the question, who holds the quality standard, excellence, right, business strategy when I'm not in the room? And immediately start to empower those people to become your AI trainers. Mm. Because like I said, right, if you have people of varying degrees of judgment that you don't trust or do trust training AI models, what they do is they just scale the bottleneck, right? They just scale the bad judgment. They just make hallucination worse for somebody who can't tell the difference between good and bad. And so if you want to start actually doing that, like make sure that that layer gets protected, and then from there you can start to design the compensation principles of saying, is that person worth two heads? Is that person worth 8X what meets expectations looks like? Google did research very early on where we realized that a great algorithms engineer, um, you know, it generates, I think it was like- At least 10X, if not more than that, right, the ads revenue of somebody who is just barely meets expectations. And so then the curve, right, in those bands went up very, very high. Somebody who is a high performer can make 8 to 10X more than, you know, somebody who is just starting in the same job family with the same exact job title because we just saw that performance, like, earn out. And it was not, um, competitive or sustainable for us to not pay them at the risk of going to our competitor with everything that they've learned in the time that they here-- were at our company. So it was always better to pay the extra money to, like, hoard the talent here so that they didn't go to a competitor or in our new ecosystem so that they didn't become a consultant charging you mon- more money than you would have paid their salary to do the same work that they used to do for you. And how many have we seen that? I've seen that happen over and over again in the nonprofit sector, right? I used to be your grant writer that you paid 60K. I am now the consultant of Grant Writers, you know, LLC with, you know, eight successful grants in my portfolio, and I charge you 150K to do this project to write grants for you. Yeah. Right. The future for sure. Yeah. All right. For, for our listeners out there, here's what I want you to take away from this conversation. Um, the AI-driven companies that figure out how to pay for impact, not hours, not titles, not, uh, out of, uh, date benchmarks and so forth like that, uh, those are the ones that are going to attract and keep the best talent in this new world. So we're not talking about the, the distanced future. This is all happening right now. And if you're someone, uh... you or someone on your team who's using AI s- uh, in a, uh, a way that Maximos has been talking here that produces the output of three to five people, and you need a compensation strategy that reflects that, you know, reach out to, to us here at CompTeam because we can definitely help you with that. If this, uh, conversation got you thinking, please subscribe to The People Strategy Forum and share this episode with your HR team, your CEO, or your board. Uh, we see the, the world changing pretty quickly, so, uh, make sure that you're in front of the game. Until next time, uh, pay people for the value that they create, not the chair they sit in. And, uh, we'll see you next time on The People Strategy Forum. All right. Take care, everyone.