Why Moveworks is a Most Loved Workplace: A talk with CEO Bhavin Shah

Hey everybody, welcome to The Leader's Show. I'm Louis Carter and we have Bhavan Shah on with us today, the CEO of Moveworks. And Bhavan is awesome. He is an entrepreneur. He grew up in Silicon Valley and he has had Moveworks for 5 years now, growing this company, amazing company with 3 other founders, and it's become a most loved workplace. It's very impressive what they do, the information and processes that they've developed to make really our lives easier inside of companies. The processes that normally take us so many hours to do, he has in Moveworks really made them truly simple so that we can have the time to do things that we need to do in our lives and in our work. Bhavan, it's great to have you here today with us.

Lou, thanks for having me. And it's great to have this conversation today. There's a lot of a lot of timely topics I think we can cover.

We can. And, you know, we have so many different things to talk about, right? We have organizational culture, how you've developed that at Moveworks and how you've become a most loved workplace, which I want to go— I want to dive into. And, you know, part of the reason, you know, that I see a lot of companies bring together a lot of companies isn't just— it's really, it's a combining or unifying vision or mission. And your folks at Moveworks do something pretty extraordinary— machine learning and AI and really retraining models and doing something that they have a unique talent to do. How does that play into how you've created your culture and choosing the right people for your team?

Yeah, it's a— look, there's a lot to unpack with a simple question like that. I think when it first comes down to our mission and how we think about the job to be done, and that is we're using conversational AI to transform how businesses work. And fundamentally, that transformation sits at this location of support. If you think about support as a kind of a fundamental need that employees have at work, they need access to tools and they need to get unstuck when they do, or they need access Once you provide that, then your employees can actually focus on the work that you hired them for, the work that they're specialists in, the work that they enjoy doing, have set out their career to become good at. And then last but not least, once they can do that work, they become, they have a sense of belonging. They have a sense of ownership that really only comes together once they're, distraction-free, they can focus on their, on their work, and then they can really act and think like an investor, like an owner, like someone who belongs at the organization. That sense of belonging really matters. So when it comes to, you know, our world and who we hire and who we bring on, of course, like any other company, we look for the best talent, the smartest talent. But there's a lot of principles and attributes that we look for in terms of the individuals, in terms of their leadership, in terms of their ability to act and operate in a fairly unknown space. We're inventing a new category. We were starting this journey about 6.5 years ago, and, you know, we were telling everyone that conversational AI was going to be You know, front and center in terms of how employees could get more work done to be more productive. Fast forward to today, I don't think anyone is denying that fact given the popularity of some of these more consumer-driven solutions like ChatGPT. So when we look at folks, we look at a variety of attributes of who to bring on. We look at their ability to work in an ambiguous environment.. So you have to act on impartial data. That's, that's number 1. Number 2, we have, you know, this mindset that no improvement is too small. And so from there, you know, we look for people who are really focused on the details, who are focused on perfection. There's a lot of perfectionists here at, at Moveworks that, that we hire. And then we also work with people who are unafraid to act on impartial information. People who have strong building and execution quotient is key. So I think these are the types of attributes that, that we look for. And, and then you have an exciting project and the rest comes together.

That's critical, isn't it? The, the execution component and, and having that kind of perfectionist detailed work, showing up on time, doing the work specifically, caring about that outcome, because in machine learning we have to annotate, we have to think about the details, we have to retrain constantly. And if, if you miss something, we can get caught with something that looks ridiculous and it makes us look terrible in the, in the, in the meantime. And so one of the things we were talking about, Pavan, was a talk I had had with the CEO of Automation Anywhere, is that when we do this, when we can train into perfection, these models and what you're doing, we will then as a society have more time to sort of surf the rings of Saturn and harness the energy of black holes. Do you have that vision too at MoveWorks, and how does that work for you?

Yeah, I think, well, let's talk about automation as a larger topic and a trend. Obviously, there's been a very big push towards, you know, digital transformation and automation across enterprises worldwide. And that is for a lot of reasons. Obviously automation helps organizations do more with less, but it's our greatest deflationary tool that we have as business leaders, right? And over the last 18 months, we've all experienced a lot of inflation, a lot of extra cost to run an organization. And so automation really becomes central. But you bring up some of these other points, which I think are, are very critical in terms of conversational AI, AI in general, and that is as we build bigger and more sophisticated systems, large language models are becoming the kind of the reference architecture to how to build a conversational, um, you know, system that can interpret and that can, can respond. But to do that, you know, you train it on data, but that's not it, right? That's, that's not the only thing that, that you have to do. You also have to annotate. You know, there's this debate around unsupervised versus supervised learning. And, you know, it turns out if you want the kind of precision and recall that you and businesses expect to have with solutions that can work out of the box, that can actually deliver real employee impact, well then you have to do a lot of annotation and you have to train. The folks at OpenAI have hired thousands of human annotators to actually provide labeled training data. And it's, it's all in the details, right? Because as soon as you're talking to a system that is aware of, you know, the knowledge or the information that you're seeking, small perturbations in the models can actually lead to pretty unsatisfying results. And so what we do a lot, and our team is very indexed on, is really looking at all of these different outcomes and looking at precision metrics and recall metrics and the curves of these sort of phenomenon to ensure that we're achieving better results with more data or it's improving over time, not regressing. So that's the detail work that has to happen in AI. And sometimes people just think it's a data problem. You just add more data and thing just takes care of itself. But new types of models are coming out on a regular basis, ensembling these techniques, looking at them, evaluating them, and then putting them in production and getting that real-time feedback is, is central.

Because we, you know, we could take these off-the-shelf models pretty easily and say, well, that's a great idea. The thing is, it may or may not work with our work, right? And I think of Hugging Face or the other kind of open, open AI models, and they're wonderfully brilliant. And I can take one and say, well, let's retrain it. Let's think about how it works for our processes, how we create our own. It sounds like you have found those people and you actively tell me about your— like, what is a standup like at Moveworks? What is your standup? You know, your Monday morning standup? What is it? What does it look like?

There's a variety of them, right? Depends on which department you're in. I think if you are on the machine learning team, you'll see a lot of calculus on the whiteboards here with, you know, figuring out how some of these models, you know, can be managed and manipulated to achieve a certain type of outcome, a certain type of response that we're looking for. But, you know, there's a couple of principles that really define our organization. And I'll just take one step back, which is culture design, organization design is not something that most— actually, I don't think anyone is born with. You don't come out of the womb knowing exactly how to do this. You can short circuit many things in life, wealth acquisition. You can, you have a company that goes from zero to infinity overnight and all of these other sorts of things that can get short circuited, but good management practices, knowing how to inspire people, knowing how to lead, being a supportive manager is something that takes time and it takes iteration. I've had the benefit of starting several companies. I am grateful for everything that we've built here at Moveworks. And this is the one, this is the one that we're gonna take and build, you know, to infinity. But the point is, is that it takes some of this sort of experience to do so. So if you're building an organization and you're trying to come up with, or you're trying to listen for, you know, key kind of advice and soundbites, I'm glad you're here. But more importantly, know that it's, it's very difficult to get it right the first time. So keep trying, keep applying, and keep developing your intuition around this and your gut around what works. But going back to, to your question, we're very indexed on what we call XQ, execution quotient. We are not an organization that is built on a lot of, you know, debate and, and, and in sort of, you know, consensus for something to happen. We allow individuals who are closest to the data and closest to the customers to make as many decisions as possible. We actually don't believe hierarchy is designed to sort of be a gatekeeper for decisions. The decisions should be made by the organization in a cross-departmental function, functional group, or, you know, individuals themselves. And that the hierarchy is really designed to help us give each other feedback and support one another. But as a result, if decisions come to me, As a CEO, I often ask the team, why? Why has it had to come to me? Why wasn't there an individual who is closer to the problem space, closer to the customer, to be able to make that choice? You know, we say something around here, which is, are we an engineering-driven culture? No. Are we a sales-driven culture? No. We're a customer-driven culture. And so to the extent that we understand our customers' needs, to the extent that we can actually get the smartest folks to look at the data and have these conversations, we will get to the best outcome. So a lot of the standups are really about people asking what got done, what did we try. And there's a theme I, I talk a lot about here, which is iteration. People will tell me, look, we figured something out. Great, have you done it 3 times? How about 15 times? Are we seeing the same insight? Is it still holding true? Or if people have an idea, I always tell them, go ahead, try it. You know, as long as it's not an irreversible decision, it's okay. And that sort of develops a strong gut and a strong intuition around what works.

You know, you had mentioned about management and leadership and iteration, and it really gets to the question in AI probability. And when you manage and lead, uh, with the understanding that iteration is the norm, uh, you get closer to a higher probability of success. And, uh, some call it make mistakes. I, I call it iteration because it's really not mistakes we're learning from, it's what we can do better in the future that we learn from. We learn from our successes even more than our mistakes. Because it's how do we change it, make it more probable, get from the 40%, 50%, 60%, 70% probability. We'll never get to 90. I don't— Alexa, Alexi, and Google, uh, Siri, they're not even close to people like 95%, right? But we can get to 60, 70. But in order to get there, we need to have the right signals, right? And, you know, if you get no signal I'm gonna need some ideas to get to a 10% signal or some signal. So iteration is extremely important in your business, and inviting how— uh, idea— not just ideas, but inviting people to keep working towards success, keep working toward what's better. How do you, how do you enable that kind of culture with within your employees where they feel comfortable with gaining iterations and going toward that success and probability and greater signals?

Yeah, look, I think speed solves many problems in life, just in general. If you make decisions quick, if you can act quickly, you will learn faster and you will be able to then, you know, get better as you're pointing out now. My machine learning team may have issues with what you said previously about 60% accuracy. They are building models with a 99% precision, you know, but it might be a 90% recall. So there's 10% of things that the model won't actually decide on, but for the stuff it can decide on, it's gonna be very, very accurate. So that's a choice that you have, which is coverage versus precision. And when it applies to culture, these are the considerations that our team is continuously making, which is, Do we solve this problem for all customers or do we solve it for these 3 customers but get it very dialed in to their specific needs, right? And so those are the types of choices. Now, when it comes to how do you create that culture, I think there has to be a culture of safety, a culture where people can propose ideas, where people can also get access to information. You know, a lot of times in organizations, information is guarded and, you know, perhaps in the, you know, defense space, that's important because you don't want everyone to know all of the secrets. But for a software organization, a SaaS company like ours, we actually want that information to be as widely disseminated as possible. So access to reports, access to earnings and data, sales data, engineering data is really central to enabling people to then decide for themselves what to do, what they can act on, what they can't. I think there is a thing that perhaps you've used as well or heard, but culture is really the sum of everyone's actions every day. It's how we treat each other. It's how we treat our customers. It's how we treat, you know, our ideas and how we think about the interactions that we have with one another. And I think when we look at that combined with this idea that people can take an idea, doesn't matter what your title is, what your role is, I think what we strive greatly to do around here is just because someone has a big title in a room doesn't mean their idea or their decision should have a greater weight than someone else who might actually have deeper insights into the particular problem. So I think that's where we kind of, you know, encourage people. We do have to police it and sort of, you know, make sure that things aren't being slowed down by bureaucracy or things of that nature. We do have a violent reaction towards that, and as a result, we sort of keep getting better.

I can see that with having especially violent results when you have brilliant people working. And this is an extraordinary place where there's funding for and cash positive for amazing people. So if, when you're not listening to those amazing people's ideas, you're essentially losing thought capital.

Yeah, yeah. And look, there are some companies that are run, um, exclusively through the brilliance of one founder. There's, there's, uh, mega-cap companies who I can think of that, that would represent that. But I think ultimately You decide what kind of organization you wanna build. And, you know, if there's anything that we've all learned from sort of managing individuals, smart people don't wanna be managed. They don't, like, they're smart. They're, they're their own free agent. They have, you know, the same brilliant thoughts and that others might have. And essentially what they want when they come to work is a vision that they agree with. A direction that everyone is heading in the same direction, you know, together on. And, you know, to then be challenged, to be given these problems that they have to go solve and go figure out. So I think that, you know, as long as you can create a culture in which there is no entitlement and there is no, kind of organizational design that is predicated on hierarchy, but that it's flat and it's transparent. You move fast, you engage people, and you hire smart people. A lot of wonderful things happen. I learned this maybe 15 years ago, but the output— when I went from an individual contributor to a manager, you learn very quickly that your success is predicated on the output of the people that report to you, not your own ideas, your own brilliance, so on and so forth. So I think, you know, as you're building even large organizations, that continues to apply.

This whole idea of output, it's an interesting one because when there is no output, one could easily blame, or one could look within and say, what can I do to help? And secondarily to all of that, is it— are there— are they the right people for the team? Because that's essential. And in the beginning, you had mentioned how you have to have people who really have the same vision, right, which is very important, and as the company, and really believe in it, right? I mean, so it's this— how are you enabling that to happen? Were you know that they're the right people for Moveworks? What do you do to ensure that they are the right people for you and your culture?

Yeah, that's a good question. So just on the point that you made leading up to this, you know, I think something that we talk a lot about is that the mindset here is one in which nothing is someone else's problem at Moveworks. So when you're interviewing someone, you're listening, right? And you're looking for how they think, how they process information, how they act. And when people complain about other departments or something was missed or something didn't come together because of this or it was late, you realize they don't have this mindset of nothing is someone else's problem. And that's something that we look for, which is how do people describe their wins and their losses? From the past and how much of that is internal reflection, how much of that is a team mindset. And I think that sort of idea of teamwork really is central to how we build the ownership that we look for. And there is a level at which people here, are unencumbered by some of the, you know, I guess fears that people will have oftentimes in making choices, making decisions. I think as a leader, as a manager, when people are in a meeting and there is a mistake that's made or an idea that's perhaps not feasible, how you react to that how you engage really determines a lot of how they're going to act the next time. And I think that's something that, you know, we are, you know, constantly thinking about. But when it comes to, you know, the, the Moveworks culture and the hiring process, you know, we do a series of conversations. We have obviously, you know, half a dozen people that interview the individual. We do what we call the Jedi interview, which helps us get to some of the more softer topics and how they view the world. Some of these operating principles that I talked about, acting on impartial information, nothing is someone else's problem. There's no detail too small and things of that nature, the openness, the transparency are cultural attributes that we look for. So sometimes that comes from hiring people from certain backgrounds, who've experienced that and they know what good looks like. Other times it's teaching people what that looks like if they're coming straight out of school or they're, you know, new to the field.

The Jedi interview. This is interesting. What— are there specific questions to the Jedi interview?

It changes depending on the role. It is looking at— it's some situational sort of examples that we give people. In terms of how they would handle things, how they, you know, how they think about their teammates, colleagues. I think that's important. How they attribute their success, how much of it is about themselves, how much of it is about others. And, you know, what they really think of becoming a part of a team. What we try and filter out in this, these conversations, Lou, is the stereotypical person who is very smart, but also very unkind and very much about their own outcomes and their own journey. And you hear it sometimes, people come to interview and they'll ask questions about the business, which is great. They get excited and we obviously need people who care about this category. It's easy to, by the way, since We're all employees. Many of us have worked at large organizations and, you know, we know how slow things can be. Takes 3 days on average to get a typical IT issue resolved. But we also look for people who are, you know, really kind of looking at how to create impact and how that impact needs to include others and are going to go about, motivating others, how they're going to go about convincing and getting consensus around certain topics. I think these things that we look for are critical for people to have as they come on board and as we filter people out.

It's what you said resonates with me about people who come on board or you interview them and, uh, the first thing that comes to their mind is their other businesses. I've had that on several occasions, and there are other desires for other lives. And I say, what my, my main thing that I say is, what are you doing here? You should go out and start that music business. Go out and start that whatever business. I encourage you to. And it's a it's a developmental conversation at that point. And, uh, so it's like, uh, you know, I didn't realize you're coming in here for therapy or counseling or career coaching. I would— if I had known, uh, would have, uh, been a different billing process. So it sounds like that happens with you too. I've always experienced it, um, too. Uh, you mentioned development because once you people come in the door who have the vision, who are ready and are thinking and ready in the movement that you're, you're leading. Um, how— what does your development process look like for those folks? Career advancement, career development— what, what do you do to help them as well?

Well, this also goes back to, you know, with the interview process. Um, we want people who are ambitious, who want to grow and develop in their career.. But one of the flags that we look for in the interview is, are they coming here only for that? Because if that's all they're looking for, which is how do I become a manager, how do I become a director, VP, et cetera, then really, you know, going back to our previous conversation, they're in it for other reasons than the overall success of this company. And what is a company? It's a collection of people who have agreed on a common pursuit. And, you know, I guess economists would say company purpose is to produce profits in solving the problems and needs of the world. So that's— combine those two, that's why we should all be at work in, you know, in our point of view. But when it comes to development, I think there is a series of sort of pieces of infrastructure that you need to develop. Understanding where people are from a leveling standpoint and creating levels and creating definitions of what those levels are and aren't. Very important, you know. And it's, and it's not just about tenure. It's not just about very kind of loose, loosely defined, you know, tasks of things. It's, it's talking about impact. And I know that in, uh, our, our leveling documentation, you know, to go from one level to the next, it's like you have to be in charge of let's say an entire product. That product has to have a certain type of impact. You have to have this certain scope. And so sometimes it's actually a constraint of the business because, you know, to go from this level, you know, 5 to 6 or something of that sort, there has to be both the opportunity that the business has for you to take on as well as your ability to be successful at that. And so, You know, when, when you're, when you're moving up, you have to be able to run multiple products that you have to exceed expectations for a certain number of quarters. Those products have to deliver certain kinds of outcomes. Well, what if there's only 2 new products this quarter and there's 6 folks who wanna lead them? Well, then, you know, others are gonna have to think about, you know, other ways that they can deliver a similar impact than, than perhaps, you know, what we've outlined. So there's a lot of discussions around that. I think we also are a company that has really focused on performance reviews since the early days. I remember when we were 6 people, Lou, and we did performance reviews and it felt completely ridiculous to do because we were all in the same room. But you know, when you give people feedback, it's the greatest leverage you have with them, on them. And it is uncomfortable. It can be difficult sometimes to hear, but it is really the best way that we all can get better at what we do. And so performance reviews, we do 'em twice a year at this point. We used to do 'em every quarter, but just at our scale, we've kind of found that twice a year is a good, good cadence.

One of the things that at Mozilla Workplace we have is something called Do Well, Do Better. There's some feedback here. And it sounds like you're of the same mind in terms of the feed-forward approach. What can we— what can you do better as a more effective leader? Is that something that you also follow?

Yeah. And, you know, I think the, you know, companies are driven by personalities and, you know, personalities are, you know, I think there's a lot of different kinds out there. The founders and I have all had the mindset that, you know, there's always a better, like, there's always something better we could be doing, you know, or we could achieve than what we just did. And we just finished our quarter. Yesterday, and it was our best quarter ever, but I'm not celebrating. There's no wine being poured and being popped. I'm just thinking now, okay, how do I keep building and how do we keep creating our impact and reaching more organizations? It's great that we are in organizations like LinkedIn and Chubb Insurance and DocuSign and Broadcom and many organizations, but What about the ones that we aren't yet in and how do we get there? So I think the growth mindset is something that you hear a lot, which is how do you get better? I always tell people there's a better version of yourself. You just gotta go have the desire to go find that person. And it extends through everything. It's how we are as professionals, but also how we are as individuals, how we treat our friends, how we treat our families, how we treat our body. And how we take care of ourselves. All of these different, um, uh, kind of aspects of improvement are ones that I think feed into each other quite like seamlessly, and they influence one another. If you get very fixated on something over here, you know, another aspect of your life just starts getting better too. And so I think that self-awareness is something that the performance reviews bring forth, but also the culture of people just giving feedback, I think, is, is, is a big part of that.

No thrashing. Number one software feedforward ever. Focus, getting to that high 95, 99, you said, level of probability. That's amazing.

For certain models, right? You, you need to have them be, because you don't want them to misinterpret. What you'd rather them do is to skip. So you can leave that particular interpretation for a human agent in our world, an HR person who can then maybe interpret something. But for the vast majority of things, these are quite mundane, routine things. So yeah, high precision is awesome.

It's awesome. I can't wait to see it. 'Cause we're talking about things like contracting, you were mentioning, right? And HR processing, things that can be done easily, right, by humans, but takes up a lot of time. And AI can— that you've done at MooseWorks solves for that.

Yeah, the frontier of AI right now that is really getting worldwide attention is language. And we've talked about computer vision and self-driving cars, you know, 5, 7 years ago. That was sort of where a lot of investment dollars were going. But now what we've seen is that there is a lot of opportunity to interpret language. And language is at the core of our species, but more importantly, it's at the core of how we get work done. And so we wait for each other to make, um, to, to read our message, to then take an action. What we've discovered, but now the world is discovering, is that using AI, we can actually have the machines interpret a good majority our customers see 50, 60, 70% of all issues completely resolved by the machine. And these aren't simple, Lou. These are like, you know, people will say things, instead of saying, "I need a new laptop," they'll say, "Hey, I've been here for 2.5 years. I think I deserve a refresh. Who can help me?" And for an AI, conversational AI bot to come back and says, "Ah, sounds like you're looking to update your primary device. I found a form." that we can fill out together, people are like, exactly, that's what I need. That's where I want to get to rather than waiting a day and a half for someone in IT or HR to kind of get back to you and say, hey, here's the process. So I think we're gonna see a new frontier over the next several years where a lot of the busy mundane work that even these support teams are trying to get off their plate, trying to make more self-serve, is really going to go through sort of this Cambrian explosion where a lot of these things will get done through machines. And then we'll focus as individuals, as support agents, as leaders on more sort of intense, complex problems. The one-off problem that occurs once a quarter, not the problem that occurs once every 7 minutes.

Absolutely. Wow. It's really been a pleasure, Bhavan. It's been a great half hour with you on The Leader's Show. Congratulations on being a Most Loved Workplace. That's where the sentiment comes in, emotion analysis of all the text comments that your employees provide to you. We're listening to that, our employee listening tools, and annotating that and understanding that you— I don't know if you realize that, but we do that as well to make sure for our Most Loved Workplaces to ensure that we know that you are really a place that people love and that you have these attributes and characteristics of people who love you. And they do with the vision that you're setting forth, with the team that you're building, the way that you co-create, the way that you respect each other. We've talked a lot about that and what you give them to develop and grow and succeed. So, Bhavan Shah, CEO of Moveworks, thank you, Moveworks. Thank you so much for joining us today and me on The Leader's Show of Newsweek.

Thank you, Lou.