Episode Transcript
[00:00:00] Monica: Welcome to UVA Data Points. I'm your host, Monica Manny. In this episode, we welcome you to the 2025 Women in Data Science Charlottesville event hosted at the University of Virginia School of Data Science with Charlottesville seeks to increase the participation of women in data science and feature outstanding women doing outstanding work. Leading the conversation is Lisa Bowers, a former executive with Genetech Roach and current director of UVA's Enterprise Studio. She is joined by our keynote speaker, Lexi Reese, CEO and co founder of Lanai Software and UVA alumna, who brings experience spanning tech giants like Google and Gusto. Drawing from their wealth of knowledge at the intersection of innovation and enterprise, Ries and Bowers share their unique perspective on how data science is shaping the future of work and innovation. From empowering the next generation of data scientists to the real world impact of AI, this fireside chat dives deep into what it means to build meaningful, transformative careers in data science.
[00:01:01] Lisa: Hello everyone. I'm Lisa. I'm the managing director of something called the Enterprise Studio here at uva. We're a new venture on grounds and our job is to help faculty, grad students and research staff create real world impact with their research. I'm here to interview and talk to my very, very dear friend Lexi, as was mentioned. So what was not mentioned is Lexi's been my friend for many, many, many years.
And I couldn't be more excited that she came to visit us here in Charlottesville today.
She was a UVA undergrad, went to Harvard Business School and then has worked at American Express, at Google, she was the COO of a company called Gusto. And now she's the CEO of Lanai Software, which is an AI company. And she's thinking about really hard things all the time that I think are completely right down the line here with the School of Data Sciences. Before I kick it off and ask Lexie some questions, I do want to thank Phil and Emma and the wonderful Emma and Emma and the wonderful team here at sds.
This is kind of becoming a second home for me a little bit at uva and I'm so happy to be here again today.
Lexi, you were a Latin American history major.
[00:02:14] Lexi: Yes.
[00:02:15] Lisa: And now you're leading an AI company.
[00:02:17] Lexi: Yes.
[00:02:17] Lisa: Explain.
[00:02:19] Lexi: Yeah, it's non obvious. I'm actually meeting with my professor from history class way back in the late 90s, Brian Owensby. And there was another really influential Latin American history professor, Tycho Braun. And when I think about UVA and what I got from the humanities, it was basically be hard on systems and Soft on people.
We were really diving into Latin American dictatorships in much of what I was doing in the last two years.
And Tycho Braun said something to the effect of, I don't expect you to like everyone that we're studying. I expect you to understand them, and I expect you to understand the construct in which they're operating in.
And I think that insight that people create systems and structures and those systems and structures can be reimagined when you assemble different people has actually been remarkably important in a career in technology.
[00:03:30] Lisa: Tell us about the kind of jobs that you've had and then explain to us about what Lanai does and what it's like.
[00:03:37] Lexi: Sure, sure. Yeah, yeah.
How many of you guys are undergrads here?
Okay. And graduate students.
Okay. And faculty and staff. Okay, great. So, you know, you always can describe your career in the rear view versus when you were doing it. So it's all gonna sound very neat and tidy, but for the undergrads in the room, I had no effing clue what I was gonna do when I got out of school. And I would say there are three chapters to the career. It was heart, head and gut. So I left here and I went and did documentary film work in Latin America again. Latin American.
Gotten connected through a UVA friend to a filmmaker. And I was told by a mentor, go get some hard skills. I didn't know what those were, but he had gone to Harvard Business School. I went to Harvard Business School, and I found the language of business to be a language of movement building. And what comes naturally to many of you in this room did not come naturally to me, which was numbers.
And the fact that if you can understand numbers, you can tell stories with numbers. And language effectively requires data. When you're in any discipline in business, whether you're in marketing or finance or product development or, of course, engineering in the more typical places you see it.
But I went to American Express because there was an incredible leader after 9 11, whose name is Ken Chenault. He was one of five black CEOs in America. And he put American Express right back. Back in the World Trade center after the events of that day. I had no idea what you did at American Express. I really. I could never have qualified for a credit card there. And I wasn't sure why they needed 150,000 people to be giving people credit cards. But he was an incredible leader, and I followed him there. And I would say I learned really that I was good at business. And I think life is about both what do you love and what are you good at? And those two things may not always coincide. Like, I love the power of documentaries to move hearts and minds. I'm not the best filmmaker. I don't love the way you guys love data science. But I trained myself to know that because I realized it was a powerful skill.
So the last chapter is gut. The last chapter is integrating. What do you love? What are you good at? What with. What do you feel is the right place for you? I found myself in tech because I had been marketing Amex to small and medium businesses, and I ended up getting recruited to Google to do the same thing when Google was 4,000 people and we had a product called AdWords that most small businesses didn't understand was a way to market themselves to people all over the globe. That was an incredible basically decade at Google, but with a semester abroad at Facebook, where I was increasingly responsible for areas of the business that have become well known today but were very kind of minuscule and sort of an idea in people's minds back then.
I'll wrap it up just because we could spend time on this in the Q and A by saying that I wanted, after Google had become very big and I was running all the advertising off of google.com to go back to something smaller. And I was the COO of a very small at the time, HR platform that was a payroll technology called Gusto. Anybody heard of Gusto? Okay, good.
Once upon a time, no hands would have gone up and participated in that. Became a unicorn tech company.
And it was incredible to be part of that journey.
I then was a glutton for punishment and said, well, actually I want to found a company. I want to found an AI company At the moment where AI is as trendy as online coupons was in the day. And this could be a terrible thing, but I decided that in this back half of my life, I want to be participating at the forefront of how businesses are making decisions about how to deploy AI, which is, I think, going to be the most consequential decision to economies and societies the world over over the next 20 years.
Lanai is a AI productivity platform that helps the world's biggest businesses see how AI is being used in their enterprise.
We real time label the risk and the use case of that AI use and then we are aiming to make people really powerful, safe, ethical users of AI.
[00:08:40] Lisa: So let's talk a little bit more about Lanai then. What are some of the things you've learned starting it? I mean, this is. You've done amazing things in your career. This is the first time you've started a company. This is your first job in an AI company.
What have you learned? And specifically about how the technology impacts people.
[00:08:59] Lexi: Oh my gosh. So first of all, there is. You can read how many of you have started a business.
Okay, so there's something called obsess about your customer. You know, obsess your obsess about your customer. And you have to find product market fit. And you actually have not started a business if you're not actually earning revenue.
And so right now I haven't started a business. I've just, I've basically raised $10 million and I'm losing most of it every day as I try to get someone to pay for the service of Lanai.
But finding product market fit is a combination of data and storytelling. So I have had probably 200 meetings over the last eight months with leaders of Fortune 5000 companies, especially CIOs and Chief Data and analytics officers, asking them, what does the world of work look like with AI in it? How are you seeing today how your employees are using AI? How are you making decisions with data, if you're using data at all, about where AI will take over certain tasks and when you'll keep humans there?
I've been asking these questions and part of entrepreneurship is really again formed here at uva, knowing what questions to ask and how to tailor those questions and know how do you do discovery?
Discovery questions are helping you inform how your product will be built.
And I am at a stage where my CTO and my chief Product officer and all my engineers and we're at a stage where we're small enough that everybody can hear those conversations. I record them if they let us be record them.
And I keep that treasure trove of interview data and use that to say what is the customer really asking for and what's the spirit of their request?
And we have now what's called the most valued, a minimum valued product in the market that a Fortune 500 company or 12 are testing and tuning and iterating. I think that it's just so materially different to have a cool idea that is interesting and nailing something that is interesting, that is urgent for a paying customer to solve and then establishing the credibility that you are the ones to solve it. And you have to have all three. You have to have. This is important, this is urgent and I believe that you can solve it. I'd say lesson number two, it's not so much a lesson as it is a reminder.
Culture eats strategy for breakfast. It eats strategy for breakfast. If I wanted to hire for the roles that I have, which are, I'm hiring 15 people. I have nine of them.
If I wanted to have people check the technical requirements on the box, I would be done hiring right now.
But I don't. I hire for values and motivation first, and role related, specific second, and very important for everybody in this room who will be in a job market.
What I'm looking for is not distinct than what most leaders, especially in tech, but not exclusively, are looking for. I want the candidate to fire and spark curiosity, urgency, empathy and drive.
And I want to see that and feel that not just in what you say, but how you say it. You come prepared.
You've done your research. I mean, for crying out loud, if you Google me, it says what questions I ask in an interview. I'm always surprised that, like, people are like, wow, that's a really hard question. I'm like, really? Because it's fucking posted on LinkedIn.
Ding.
No, but I think that, you know, this will get to what Lisa and I are going to talk to. But to be great at your job technically right now, technical skills will come and go. Your curiosity and your clarity of communicating your unique capability to deliver value to a company that is never going to go away.
[00:13:20] Lisa: I just want to shine a light on a couple of things Lexi said. One is she's hiring, just saying. Did you see the difference between the nine people she has and the 15 people that she needs to hire? So just know that.
But secondly, I think this point about seeking to understand is really relevant for everybody in the room. I mean, I'm thinking like, say you graduate from this program and you go out into your first job and someone gives you a project to do. You can either sort of imagine what you think it is and then start or ask a lot of questions, really seeking to understand what your customer needs. Whether you're the CEO of an AI company or whether you're an incoming analyst in a large corporation who's been given a task, seeking to understand the purpose of that task and the need of the client, whoever the client is, I think is a great learning for all of us.
Okay, so you've been basically at the sharp tip of the spear for technology in so many ways, like your time at Google, your time at Gusto, and now here.
What excites you the most about what data scientists are doing today?
[00:14:24] Lexi: Oh my gosh.
[00:14:25] Lisa: What Jazz is you.
[00:14:26] Lexi: You guys are so cool. First of all, I want to know all about what you're doing.
Wid's founder Here I love in AI the notion of a triple check system.
Has anybody heard about that?
So you know, AI, generative AI in particular, is, is eating a lot of data and you want to check to make sure that you understand what is this big model that eats a lot of data, what data is it eating and what data is it obscuring?
Because whatever the model is trained on is going to influence what the outputs of that model are.
So looking at data before really deeply looking at data during running experiments on the model and what are the outputs of the model and then running experiments after periodically in a regular way to say is this because if you think about large language models or fine tuned small models, it is just like a person, you know, it's a person who's getting smarter based on whatever you're putting in it as well. So it's never static, you're never done as a data scientist checking whether this person, and I think of AI more as a teammate than a tool, is behaving in the way that we want it to. So I like this notion of there being data science discipline of a triple check on models themselves.
I also like counterfactual testing. So making sure that while your version of ChatGPT and my version of ChatGPT, if we ask ChatGPT this the exact same question, we are not going to get the exact same responses because it will. Your chatgpt will know you and know what you're looking for and what I'm looking for differently.
But if on a material basis the answers become inequitably distributed and information is given to some and not given to others, that could inform large decisions.
That's counterfactual checking to make sure that the model behaves equitably, no matter who is using it. I'm really psyched about that field of data science too.
[00:16:51] Lisa: Excellent. Thank you, Lexi. Okay, so let's talk about our audience here.
Many folks here are wondering how to make their mark. Like how do they get started, how do they grow, how do they design a career? What are the opportunities that you see for the next generation, for this generation of data scientists?
[00:17:09] Lexi: Yeah, first of all, I didn't realize the connection to Stanford. So I live right near Stanford. A lot of the people in that video are friends who we've known each other 30 years and they were all just like you guys. Like most of the people there are coming from grade schools who had similarly, maybe some people were very well directed, but most people were just figuring it out, you know, like bumping into things they loved and bumping things they didn't love.
But a couple things stand out in particular, which is again, they all evidenced curiosity and openness to feedback and particularly in the data science field. Susan Wojcicki, who is the CEO of YouTube, who died last year, and it was a very good friend, she was an incredible technical leader.
She could not communicate for shit.
When, when she gave a speech like, I was leading an advertising business that she, it was her engineering that we were representing and she got on stage and I was like, oh my God.
But she was Susan Wojcicki. The Google started in her garage. And like, I didn't want to be offensive, but she asked me how she did and I was like, do you want the answer or do you want what? I think most people would tell you as the answer said, I want the answer. I said, susan, no one understood a word you said honestly. And I think you could do so much better. But you have to remember that in general, when you're speaking to a large crowd, you've got to translate this technical concept into stories, metaphors, language that people are going to remember because you're not communicating to a technical audience. In large part, you're communicating to be understood and to make people feel, think and do something as a result of what you said. So for everybody who's speaking ever, what do you want people to feel, what do you want people to think, and what do you want people to do afterwards? And I think in data science in particular, you all have the opportunity to be these incredible detectives. You're looking at pools of data and you're spotting patterns and you're seeking answers to a question.
And that question is not for question sake. It has a purpose. So what question are you really seeking an answer to and what are the patterns telling you? And then how do you communicate that back in a really compelling way?
[00:19:43] Lisa: Thank you.
Okay, so I have to ask a question about entrepreneurship because that's my job.
Tell me about the kinds of ideas that Data Sciences might launch into a company. Like, how could Data Science, even from uva, translate into entrepreneurial potential?
[00:20:03] Lexi: Oh my gosh. I mean, so much. First of all, like, let's not see entrepreneurship through rose colored glasses. Entrepreneurship is very, very challenging in ways that nobody talks about. When you just hear like, this startup went from zero to a billion dollars, I mean, you're plugging, you're plugging things in.
You're humbling yourself. Do you know how many people I have reached out to for machine learning jobs in Silicon Valley? I'll tell you how many? 2,833.
And do you know how many? Swipe right on my request. Not many. It is very humbling to seek everything that you need for entrepreneurship, which is the idea is easy, the execution is hard. For execution, you need money, you need people, and you need relentless energy that will not stop against a thousand no's. What was your question?
[00:20:58] Lisa: Well, that's.
This is the visceral, sort of emotional reaction to the word entrepreneurship.
[00:21:04] Lexi: Yes. Oh, yeah. What are you guys.
[00:21:06] Lisa: What does that mean to these people?
[00:21:08] Lexi: Yeah, well, what that means is there's lots of ways to be entrepreneurial which do not correspond with necessarily starting your business. I would never discourage someone from starting a business. I would always encourage someone for being entrepreneurial in whatever you're doing. I think data scientists by nature are this way insofar as you are willing to look at data in different ways than most people are.
And you, again, have this crazy ability to see things in a data set that no one else can.
I think when you are really amazing in entrepreneurial sense, you're not only seeking the question that someone tells you to find an answer to, you're asking the next question, and what if this.
And what if this, Like, I'll give you an example for gusto. Basically every business has to get, teach, and grow customers. Like any business, we have to find them, we have to get them on our platform, and we have to grow our relationship with them over time. And if you're in business, most of the questions that you are going to be asked are going to be about, where are we breaking in the funnel? There's a pool of people we could be getting as customers. Who are they?
Why aren't they all visiting our website or coming to find us?
And then if we touch them in some way, if I have a sales conversation or a marketing conversation, why don't all of them become customers?
I think sometimes people who are in data science get too specific about finding the answer to the question they've asked instead of stepping back to say, oh, what I'm really looking for is an unlock on how customers can be found for this.
Or I'm really looking for an unlock on how do we make sure once customers get here they can stay.
So just the best data scientists that I see translate into real entrepreneurs are able to, again, ask the question behind the question.
[00:23:20] Lisa: Excellent.
Okay, so we talked about how the question behind the question can help with entrepreneurial thinking across a lot of different characteristics. Tell us about how young data sciences, how young data scientists can drive change in Established organizations. So when you're in an organization and you see something that's an issue or you want to actually make a name for yourself, what could you do?
[00:23:48] Lexi: Yeah, I mean, so nonprofit, for profit, public partnership, whatever it is, there is a vision of that organization, that organization exists to bring about a world that doesn't exist today.
What is the vision? That organization has a mission.
Their product or service will create that world by doing something. Something.
What is that something?
And what is the plan to get to that future state as described by the CEO or the leader of that organization?
Know that plan, know those priorities, and know your part in it.
Again, people don't lose agency. As much as they give it away, they give it away. Like, I will often talk to young people in any career and they'll say, well, they just don't understand how hard this is and they don't get it. And if only they. And oftentimes I'm the they in this situation. I'm like, well, what don't I understand?
Explain it to me. What do you see that I don't see?
And I am always willing to listen to that. But I'm surprised sometimes by how little effort people have tried to do the reverse, especially in this generation.
In a world where we grow up on social media, like calling each other out, you know, it's like we've built a story about some figure who we've never met and we have minimal information about, but we've climbed a ladder of inference with that minimal data to have a full thesis about who they are and what motivates them. And we're angry. We're like, at our phone and we're angry and we're like, we're gonna say the thing. We're gonna do that thing that doesn't help anybody.
We have so many other options to call people in, to call people on the on, to be able to engage in a real dialogue. And my advice for folks early in their career is just work hard because you are unique in such expertise to understand the perspective of anyone you're questioning before you go too far down the line of thinking you have an answer about what they're really after.
[00:26:20] Lisa: Thank you for that.
[00:26:21] Lexi: I'm so off script.
[00:26:23] Lisa: I love it. I'm like, I don't know. Are you serious? That's okay.
[00:26:26] Lisa: I'm tracking. I know where you are always.
Let's talk about advice. What advice would you give to women starting their careers in data science today?
[00:26:38] Lexi: Yeah, I think for women, for anybody who's here, I would say your perspective is your superpower. And so you want to join organizations. First of all, despite what you hear, there is too little talent going after too many jobs for the jobs that most likely many of you in this room will want to have good data scientists who have stretched themselves in the humanities as well, who are clear communicators and generous thinkers and engage with enthusiasm that we need a lot of those across the board in every field.
So that's number one. Number two, I would say when you get to a place, don't give your agency away.
You know you've been invited to any table that you're at.
Ask hard questions.
Ask hard questions, seek clarity. And again, don't wait to be told to do something. If you see that there is something to be done.
If you see that there's really a problem, I guarantee your leadership team is saying it to you in some way at every all hands or in their ask me anythings or what have you. They're telling you what's on their mind. They're worried about something. How can you help them solve those problems? And when you start to engage that way, like work hard to solve problems before you're asked to solve problems, that's how you become the person that they cannot live without.
I can name who these people are. For me in data science, Becca, Sarah, Daniel, these are people who said, Lexi, you've been talking about this issue. I've found the data that helps us crack the code on why something is not working. So be that person.
[00:28:35] Lisa: The word that comes to mind, Lexi, is ferocious.
Like starting in any job, being ferocious about understanding what's going on and contributing in whatever way it is, as opposed to sort of waiting is what I'm hearing from you, like bring ferocity to the table, right? That like every day. Where is your contribution? How are you helping lift, helping lift this organization? And hopefully it's an organization that shares your values and the things that you care about.
[00:29:05] Lexi: Yep.
[00:29:06] Lisa: Okay.
Talk about disparities in tech representation. Okay, so you've written so many of you. If you think, if you liked this show, Lexi has a substack blog that I suggest everybody read every day.
And you've written in it about how representation in the tech industry generally is only a starting point and that there's so many other steps to make sure that we have all the voices with equal ferocity, contributing and lifting.
So talk to us about the relationship between representation and inclusion and how that creates high performing teams.
[00:29:46] Lexi: Yeah, sure. So this is, we're in the school of data science and we'll use some data here. And we're in the U.S. so I'm going to use some data in the U.S.
there's a lot of controversy about the words that I can and can't use on this topic right now. And I will be very fact based in business and in governments. In the United States, there is a statistical anomaly of an overrepresentation of white men in rooms of power where product decisions have been made and where policy decisions have been made. And if you are in the technical community or in public service, why is that important for you to understand and want to address?
Because the data suggests that teams that are both representative by gender, race and other forms of background, including ideological, perform at much higher rates. How do I know that? Based on longitudinal studies on return on invested capital of these businesses and based on employee engagement, that is two to five times higher for businesses that represent heterogeneity in their leadership team and at every level versus homogeneity.
So what I do not like about how language has been weaponized here is that it neglects the very specific data that we have to say. When you hire for homogeneity, you hire for mediocrity.
But if you are in a position where you are largely a homogeneous leadership team and you hire to correct for representation, but you don't change anything in the organization about how you include other people's perspectives when they are there, you actually perform worse than if you had stayed homogeneous from the get go.
Why is that?
We have probably all felt this in one way or the other, because when you feel like you are the box that has been checked and you enter a room and that room behaves as if you never entered it, and they're talking in a way that talks about things they've done or seen or what have you, and no one takes the very easy step to empathize. What your background is, where have you come from?
That person first gets sad and then gets angry and then they leave.
And it's that simple.
And what does that mean? Very practically, it's not that hard. We substitute an approach that many of us in the US have been taught, which is the golden rule, treat everybody as you want to be treated. We substitute that leadership approach with a different approach, which is called the Platinum rule.
Treat people as they want to be treated, which just means that I take the time to understand who they are, what they come from and what they care about. And then I alter in small ways, not lowering standards. I have very high standards for everyone and I have very deep care for everyone. But I'm coaching people situationally based on what brings out the best of them.
[00:33:10] Lisa: Thank you, Lexi. So I have several other pages of questions, but before I jump into more, I wanted to turn it over to the room to see if anybody has anything that they'd like to ask Lexi.
[00:33:21] Attendee: I'm here with a sophomore, hoping to influence her to come to uva, but.
[00:33:25] Lexi: So you both is this year, Sophomore. Okay, well, let's all do that.
[00:33:31] Attendee: She's going to kill me in the car, you know.
[00:33:33] Lisa: Right.
[00:33:33] Lexi: Yeah. That's okay. That's okay.
[00:33:35] Attendee: Serious question. Given everything that's going on in the US Right now, do you feel, feel that we were perceived as the country where higher education happened, where brains and all these innovation come out? Right. And with the use of data and manipulation of data for bad, what do you think our country will look like five years from now in terms of data science, innovation, et cetera? And if it's veering from your topic, then you can skip me, but love it. You guys are so fascinated, and I want to get your thoughts and maybe a little bit of hope from you.
[00:34:07] Lisa: We're full of hope. That's all we're about. You want to start?
[00:34:10] Lexi: You go, yeah, well, so what will our country look like five years from now? I mean, honestly, this is why I come to universities. I mean, again, I went to uva. I would say I am so impressed.
I like UVA more now than when I went here. I am so impressed by how this school has navigated the events. Were you.
Is that important? I mean, do we. 1996. 1996.
Okay, all right, okay. But we're like, same, same over here.
I think that this school's ability to add STEM curriculum but continue to focus on how important the humanities are is representative of where I hope this country will continue to go. Because I think there's a lot of bluster in adult corridors.
Whereas people in this room and in this university and universities the world over are really just trying to navigate to a world that they're excited about that is sustainable. And I don't think that this formula is hard. We need to make work work. We need to make the government work and we need to help families to work.
And we know how to do this. We have a lot of adults blocking and creating traffic, but we can make that happen. If good people, like everybody here, makes decisions about who they work with, that isn't just based on the dollars that you're going to get from folks, but you pay attention to what are the values of the organizations that you're going to join, who are the people leading them?
Don't do the justification of, yes, this is a terrible human being who represents everything that I do not stand for, but I'm gonna work for them and make a million dollars and I'll wash it away on the back end. No, don't do it.
Your superpower is your talent. And where you choose to spend your talent and your time and your treasure matters.
And like we started at the beginning, people create systems and people change systems. In a capitalist economy, your talent contributed in an organization is your biggest currency next to your consumer dollars. So choose wisely. And I think this generation will. And that's what's going to make for a future that is even better than any that we've seen before.
[00:36:44] Lisa: I'll just add that I agree two things. One is find where the good people are. And actually Lexi and I have discussed this. Lexi and I have both driven our careers to generate as much power as we possibly can have so that we are awesome and we're going to make awesomeness and have the power to do so. And I think anybody that is in this room has exactly the same capability.
So we are here because we did not settle for small and that it's not easy and you know, we're full of bumps and bruises and scars and all those things. But power is what is making the world great or not great. And so go and get that power, everybody, and then deploy it for, for the good, just like the school of data science does.
[00:37:31] Lexi: I don't know if this is on, but Hi, I'm Jade Preston. I just was curious how did you go about raising the funds to start your company?
So this is a story about relationships as well. My co founder was the CTO of a company called VMware which is a large tech company. And he created great comp sci data scientists. Stanford, we'll forgive him for that.
He has been an investor for the last 12 years and we overlapped while I was doing a residency at a venture capital firm and I had an idea that was like lanai. He had an idea coming from a very different angle. His angle was in the cloud transformation. When we were moving from on prem to clouds, there was technology created that was called observability tools that would tell enterprises as you're moving to Amazon cloud or Microsoft cloud or Google cloud, like this is how performant the cloud is. And he was like there's going to be something like that for AI.
And we connected the dots and I said that's a great product wedge, but this is really about democratizing the capability of everyone to be a safe AI power user. And the businesses that get ahead and stay ahead will see humans using AI, but we will use that observability layer in order to help them inform future decisions on how they'll deploy AI versus humans. Like, what are the fit for purpose reasons to use AI versus humans? I say that because, one, I had a relationship with him. Two, he now works at a venture fund. They put the first money in and then I'm in a seed stage. So we didn't do a process.
I went to people that I know very well who actually run funds that are made of operators. And so one is the Black Angel Group, which is black executives across the valley who started a venture firm. I worked with one of the co founders there at Google and the other is another women's group in tech and explained what the idea is and they were backing the idea because it has huge commercial potential.
But really at this stage, people are backing you, the founder, like, do they believe in you? Because the very real probability is that the idea that I pitch today is not going to be the one that is the one that is the product that we take to market.
So they're looking for people that are resilient, that are credible, that know how to hire, and that's how I raise the funds.
[00:40:17] Attendee: Hi, my name is Julia. I just want to say thank you so much for speaking to us all.
I am really passionate about Latin American studies. As you said, you were talking about your passion for Latin American history. And so I just wanted to know a little bit about how you continue that passion as you work and if you have any advice about people who have similar passions and about merging data science with those passions together.
[00:40:45] Lexi: Yeah. Have you taken a class with Tico Brown or Brian Owensby? Do you know those names?
Oh, okay.
[00:40:53] Lisa: Welcome.
[00:40:53] Lexi: All right, well, welcome. I don't know if, I mean, if you can like cross pollinate. I don't know if that happens. But Tico Brown has something that he started in Latin American historian, started in Covid, where he was like, write it down. Journal, journal, journal. Write it down. Every day, everybody. Journal. We are living through such incredible times, Such incredible times. It can be frustrating, it can be energizing, what have you, your diary of your observations of the world right now. And if you're interested in history, I find patterns through my writing and I write every weekend, as Lisa said.
And writing is the way that I integrate the things that I care about and am passionate about.
[00:41:44] Attendee: So I graduated about a decade ago and have had my career largely in data science and analytics. Now I'm starting my own AI company and enjoying the journey greatly.
I want to kind of go back to what you were talking about in some ways and talk a little bit more about integrity.
Many times it can be the case that people in analytics or data science roles are pressured to misrepresent numbers. Is that something you've seen and do you have a recommendation about how one early in their career could respond to seeing this sort of practice?
[00:42:22] Lexi: Lise, do you want to go? You go first.
[00:42:25] Lisa: And I'll add, if there's anything of.
[00:42:26] Lexi: Value, I mean, honestly, if it's really that blatant, if it is misrepresentation, again, this is where the five W's come in. Why? For whom?
What's the purpose? When do you need it?
Why? Again, ask a lot of questions, assume positive intent.
And if upon hearing those answers and your response is, well, I would love to support you, and the data actually does say something different, if they're asking you to compromise that value, it doesn't matter what the circumstances leave. Like, your integrity is too. Is too high. But I do think that sometimes that is a lost in translation moment where what feels like an integrity issue to you upon further questioning might be like a miscommunication, but we can talk more specifically. It sounds like there's more. There's more to it than that, but honestly, there just isn't.
Again, you guys are in rarefied air.
You're in rarefied air. People who tell stories with numbers, who can see numbers, women who tell stories with numbers in particular, really hard to find.
So you have so much more power than you might think as it relates to where you provide your services and with whom and to whom.
So you can feel small in spaces where certain people use bravado and bless her to make themselves big. But just take a deep yoga breath, take up the space and say, thank you, but no thank you.
[00:44:09] Lisa: Nothing to add there. That's perfect.
Next.
[00:44:12] Attendee: All right, so my name is Erica. My question is. Because I know you're talking about your company, and I feel like, like, for me, it's kind of a hard sell. Like, we want to be, like, I want to be efficient, and I think that AI is perfect for that. Like you said, like, when it fits and things like that. But like, how do you. When your clients. Do you ever have any clients that are like, well, yeah, that'd be great if they were efficient, but now what are we going to like basically what time suck can we put them into to justify how much we are paying them? Because I feel like in more traditional industries that's kind of like where you kind of lose agency coming in. Like, I don't want to say as a younger person, but just a person that's more open to. Okay, like we've already driven this as efficient as possible, so can we change the nature of what we do? So maybe we're doing more thinking work and even if we're sitting and we're thinking, we're still working. So kind of how do you navigate that space?
[00:45:05] Lexi: Well, here's, here's what I think and I'm going to just stand because my back is hurting. I.
If, if history, if past is precedent, what AI is, is a tool for automating jobs. And it will, despite lots of lipstick on the pig, replace workers. Because AI can do call center work, it can sell, it can do first level coding, first level legal and analytics, it can do consulting. Like I have nine people, I feel like I have 15 because I have an extra seller, an extra marketer, an extra coder with AI. That is true.
The reason that I wanted to start an enterprise analytics play that shines a light on what does AI do and what do humans do.
So I'll give you an example with Lanai. You could go to Walmart and Walmart might say, I have an idea of how we should use a set of AI tools to sell more efficiently and effectively. If Walmart is a customer of mine, I'm like, that's a great idea.
Before we roll that out across your entire workforce.
Because they have a quarterly goal to hit and what that will look like in the quarter is they will have cut their biggest cost, which is labor costs, and that profitability will drop to the bottom line. Before you roll that out, let me tell you what could happen, which is that short term profitability bump actually is really leading to sales that are low calorie, which means you get them, but those are customers that churn later because they didn't have a good experience. They bought now, but they didn't become a loyal customer. So why don't we run an AB test that says let's look at a population using these AI tools to do this experiment. But then watch the population of customers touched by that and the population touched by, by people who were only using the tools they were previously. And you can split it in many different ways and do longer term studies on the full impact of both the cost structure but also the lagging Indicators of success, which is what happened to the customer. So I wanted to do that. That's a little different than your question because I think that way we're helping people do, which contrary to many people's belief, knowing a lot of CEOs of these companies, they're not a hole. Some are. But many people are leading large businesses because they like people and they want to use people to their best and highest use purpose.
And so your second question or your actual question was what do you do with all the efficiency and how do people make the decision of okay, now that we have our call center automated, their low level work that call center. Now the smart employers are going to think about cross training them to be product development people. Because who knows better how to design a product that really meets the customer needs than the people who've been on the front line with the customers for the past six years?
And my biggest issue of not having that rotational program in the past was that I couldn't train customer service people on the skills they needed to be product development people. But now with AI, I can.
And so basically I'm trying to create what I know people have sketched out in their mind as an ecosystem that makes those pathways to prosperity easier for companies to implement and people to attain.
[00:48:57] Lisa: Who's next?
[00:48:58] Lexi: Hello.
[00:48:59] Lisa: Oh, hi.
[00:49:00] Attendee: Yeah, nice to meet you. Thank you so much for your time. Lexi, I just wanted to touch on something that you mentioned early on in the talk about like figuring out what you love and also what you're good at. And you mentioned that data science was something that you kind of had to learn.
So could you speak a little bit to how you taught yourself data science and AI and how you kind of stayed true to the business problem while still being technically informed?
[00:49:27] Lexi: I mean, honestly, I think when I was at Amex and Google, it was just so embedded in the company where I was.
I didn't really need to understand data science until I got to gusto and was really faced with how do we grow a business from $10 million to $100 million.
And when you're selling to small and medium businesses or you're selling anything to consumers, it's such a large data set of understanding who you're selling to. And we really. I was responsible for the whole life cycle of the business. So anything not engineering I was responsible for. And I started to really befriend the data scientists because they became the people that were translating the questions I had into answers.
And the back and forth that was fun, was that's a good Answer, but I have no idea how to action that.
So let's just talk about segmentation as an example.
Lots of data scientists come up with cool ideas for customer segments. And Lisa knows a lot about this, too. And Lisa, by the way, who runs the Venture Studio, which is for grad students and faculty members to get their ideas off the ground and potentially into a business, is a genius technical and business leader herself, and was at Genentech forever, knows a lot about data in the context of healthcare and reproductive health as well. So you have a person who lives in Charlottesville who can answer a lot.
[00:51:10] Lisa: Of these questions on go another time, Continue.
[00:51:13] Lexi: But, for example, we really needed to figure out who was Gusto going to be good for.
And I needed that to be actionable in the form of I needed to be able to know the segment and find them online and be able to find them online with tools at my disposal. What are those tools? It's online marketing. So I need to be able to actually map the segments of who is going to be good to Gusto to who we might find. And then we needed to create a website in a way that we could see, okay, someone got here, but they bounced right off.
Why didn't they stay?
I find that the best Data science were able to take that simple question of, wow, we've created this product, it does the following. At least that's what we say it does on the box. It should be so much more compelling than we were seeing in our growth rates. And I partnered with Data Science on effectively that research question, and that enabled us to then focus, because the answer to that question informs, where do we focus? Do we focus on repositioning the company?
Is Gusto the payroll platform better than Gusto the people platform? Or should it be called Gusto, the HR information system, or what have you? Like, those words are actually testable and informed.
A lot of energy versus spending a lot of money on, you know, salespeople at the bottom of the funnel to talk to people that have gone through our entire product flow. I think that's a long way to say I trained myself, but really, data scientists trained me because they were able to say without making me feel terrible, that question is way too big.
We're going to have to subdivide that question into smaller parts. And Lexi, you don't need to have all the data, but you need to give me your instinct of, like, I only have so much time so I can spend my energy researching this or researching that.
Where do you think it's most important and so it was just, it was an ongoing relationship.
[00:53:29] Attendee: Hi, thank you for the inspired talk.
I have one question. Like you talk about the values each company should have as a CEO and founder. How you start putting this value, is it like representing you or something? You inspire this company present, and when the company get bigger, his value should change based on your thoughts or something else.
[00:53:58] Lexi: So I don't think there's something different between personal values and professional values. And this is a benefit of having started the company.
The values that we have are my values. And there's four, there's four of them.
I talked about them at a high level, but it's curiosity, community, commitment and clarity.
And we have, when we interview, those are defined very specifically, like what does it mean to be curious? And what's a derailer behavior? Like what does it mean, mean to not be curious? Because sometimes these words, it's like motherhood and apple pie, nobody can disagree with them. But I really wanted to use 30 years of operating to say, if culture kills companies more than bad products kill companies, why is that? Well, let's just take curiosity for example. To be curious about your colleagues and be curious about your customers manifests in an always growth mindset that's not fixed.
And a derailer to that behavior is when people center a story on themselves when it's good, but anytime it's bad, they center on others.
So in interviews, if I say, what's the worst thing you've ever done professionally? What's the biggest yard sale you've ever had? Like it was a terrible. You cringe thinking about it because you screwed up so much.
And they answer that question by saying how someone else screwed up so much, you know, which happens a lot. That's an example of how we train people how to interview to hear for the values that we have. And we have a specific interview on values that I give, then we train people on the values and we've created norms and in the company that reinforce those values. So community as an example, every conference room is named after great teams. And some of those teams have failed epically and then come back to win because we believe in teams and community, brilliance over individual brilliance.
So values become expectations, become norms.
And then as we get culture carriers, the first class of culture people that we hire, I trained on these rituals now that first class trains the second class. And you start to invite people to be part of what we call the lanai way, the human systems that guide our culture, where values are the coded DNA.
[00:56:38] Attendee: Hi, I'm Ainsley. I Know, many of us are kind of like young professionals or new professionals in this room. I was wondering, kind of like, what questions should we be asking? How should we truly assess the values of a leader or a company and, like, what their vision, mission and plan really, truly is?
[00:56:56] Lexi: Ask them one.
I mean, anytime you're interviewing, I think that's a great question. What's the vision of the company when I empower humans to do the extraordinary in the age of AI mission, to discover, secure and accelerate human AI usage across the enterprise. Vision, mission, values. I just named them.
And what's the plan? The plan is we need 21,000 people being empowered by Lanai, meaning we have a browser extension on 21,000 laptops in six months. And here's everybody's part to do that. Everybody in the company has that.
So I would say if you're interviewing, it's your right to ask that. I would also say most of this is available on a website. If not, if not, exact words, just for every interview you have for a company, keep a cloud or a chatgpt or your LLM of choice. Perplexity is a good one, too. Keep a GPT open for that company and make sure that you have uncovered as much research about those questions as you possibly can online.
And then you're going in with a trust but verify. Hey, I saw in the earnings announcement the leader referred to the vision and mission and the plan this way. Is that still right? What is the role of data science in helping with that, that type of thing?
[00:58:31] Lisa: I'd also say ask people who've worked there, like, your network is so powerful. Right? One of the things that Lexi talked about is she raised $10 million predominantly because she had a great network. Right? Like people she knew who trusted her.
She can also pick up the phone and say, I'm thinking about working with this founder. Anybody else worked with. Do you know anyone who's worked with this founder? What was it like to work there?
So that would just be another add to that.
[00:58:54] Lexi: Thank you so much for coming in to talking today. I'm Sierra. You mentioned in undergrad, like, still figuring out what you wanted to do. You were still like, where am I going with this? Or, what advice do you have for someone who's really passionate about what they're learning, but they're not sure where they're going with that yet.
[00:59:10] Lexi: First of all, just enjoy that question. Like Rilke says, live the questions. Just enjoy it. You don't need to know.
You don't need to know. This world that manufactures like College is about getting a job. Is about doing something. Is about doing something. Like, no, it isn't.
Like life is so much more than that.
Take classes that really spark your curiosity. Engage deeply in those classes with your professors, with your teachers. Not to get a grade, but to learn that thing. This is something that they're passionate about. It's so great to be surrounded by that. And then as it relates to jobs, yes, I do think you should trust try out different jobs. Some jobs pay you, some jobs don't. But don't overthink it. Like what are you curious about? I was curious about documentary filmmaking. I went to Latin America, I got paid in a sweater. I mean this woman basically. But that was okay. Cause at that point I could hustle. I found a side gig doing.
I'll tell you about that story later. But there are ways to make money that don't necessarily have to be what you are pursuing your curiosity in and then just pay attention to what lights you up. I mean, Harvard Business School cut off my right arm before I thought I was going to Harvard Business School. Why did I go there? Because someone took an interest in me and said, this is a really good place. I could see you really liking it. And then I didn't prejudge it. I sort of was judgy. But then I looked, I was like, wow, this cool space, cool people. I'm going to learn a lot here.
Later in life you're going to have to be disciplined. You're going to have responsibilities, you're going to have to like people, problem space and life design.
That life design is going to change in chapters of your life. And at some point you're going to have to say, I'm going to do something I don't like either because I really have to learn this, like doing finance. I didn't love finance, but I was like, this is a stamp on my corporate passport that will help me ultimately be able to give back to a world that I care about. I never want to be talked out of the table or under a room because someone knows more about finance than me. So just like know and endure the pain that you might have to because you have a longer term plan nestled loosely in your head and you'll be so much better off than if you say, well, I've got to go work for McKinsey for two years and then I'm going to go back to business school and then I'm going to do I banking. And then I mean those people, I can tell them when I'm interviewing them too. And I'm like, this world is so much rough and tumble and I need resilient, engaged, enthusiastic learners more than I need need robots who have been trained according to what they thought the world was expecting of them.
[01:02:14] Monica: Thanks for listening to this episode of UVA Data Points. Want to hear more from our Women in Data Science event? Check out our YouTube channel, UVA School of Data Science to experience more. And if you're enjoying UVA Data Points, be sure to give us a rating and review wherever you listen to podcasts. We'll be back soon with another conversation about the world of data science.
