Gender Data & The Future Of Women-Centered AI — With Shubhi Rao of Uplevyl

AI doesn’t fail women because it’s “bad at math.” It fails when the data behind it leaves women out, flattening real lives into a default user that never had to navigate widowhood, caregiving, fragmented healthcare, or high-stakes rights decisions.

I sit down with Shubhi Rao, founder and CEO of Uplevyl, to unpack what gender data actually means in practice and why women-centered AI can’t treat women as one monolithic group. Shubhi shares how her path through software engineering, finance, and executive leadership led to a simple but uncomfortable question: if AI is powered by datasets and algorithms, where are the datasets that reflect women’s lived experiences across life stage, culture, and circumstance?

From the great wealth transfer and the hidden complexity of widowhood to the need for trauma-informed support for survivors seeking workplace rights, we explore what trustworthy AI looks like when the stakes are real. We also get concrete about privacy and safety: anonymity, minimal data retention, and clear boundaries on whether user interactions are ever used to retrain models. And we zoom out to the opportunity for women inside companies today, from process reengineering to board-level governance and risk, with practical questions to ask about sourcing, testing, accountability, and who benefits.

If you want ethical AI, responsible AI, and products that actually work for women, this conversation offers a blueprint for building, not just reacting. Subscribe to She Builds with AI, share the episode with a woman who needs a nudge to claim her seat at the table, and leave a rating or review so more listeners find these stories.

 ✨ Stay Connected with Shubhi Rao / Uplevyl:

🔗 LinkedIn Shubhi Rao: https://www.linkedin.com/in/shubhirao/
🔗 LinkedIn Uplevyl: https://www.linkedin.com/company/uplevyl/
🔗 Instagram Uplevyl: https://www.instagram.com/uplevyl/

Transcript

Julia Lach (00:00)
Hello and welcome back to She Builds With AI, Women Shaping the Future. I'm Julia Lach, and today I am joined by Shubhi Rao, founder and CEO of Uplevyl an AI company building secure women-centered intelligence hubs that turn authoritative data into actionable insights, collaboration tools, and learning experiences. Shubhi brings a rare combination into this conversation.

She started her career as a software engineer, moved into finance and senior executive leadership, held roles at Ford, Tesco PWC and Alphabet, and later served as CFO and COO of DOSH before its exit to Cardolytics.

With Uplevyl Shubhi is asking one of the most important questions in AI right now. What happens when the data behind intelligent systems does not fully reflect women's lives, choices, and transitions? In this conversation, we talk about gender, data, women-centered AI, trust, privacy, and why women should not only critique AI from the sidelines, but help build and lead what comes.

Next. I am very excited for this conversation. Shubhi welcome to the show.

Shubhi Rao, Uplevyl (01:15)
Thank you. A pleasure to be here.

Julia Lach (01:17)
Let's dive right into it. Shubhi your career brings together software engineering, finance, treasury leadership, corporate operating roles, and now AI entrepreneurship. When you look back, how do these different chapters connect to the company you're building today?

Shubhi Rao, Uplevyl (01:35)
Well, you know, on the surface, they feel like very discrete sets of experiences, but there is a common thread across all of them, and that is numbers. So I've always been a numbers girl. I've always enjoyed whether it was hard sciences or my engineering classes or later on finance or the various applications of bringing finance and technology together through fintech applications, running just pure treasury international.

finance, etc., have always been exciting and interesting to me because in the end they're all just different applications of using numbers. So I would say that's been the common denominator. And now when you think about what I'm doing with Uplevyl is frankly an amalgamation of a lot of those experiences. But again, what is AI, right? In the end it's again another numerical application of data and technology.

Julia Lach (02:25)
That is right, that is right. Thank you, Shubhi. I really love starting here because it just shows that AI leadership can have so many different layers, silhouettes, and does not always have, the straight technical path. And I think that connects just beautifully to the moment when you just started seeing the data question behind the technology. Shubhi, was there a specific moment when you realized?

That the question was not only what can AI do, but also whose data and whose lived experience is missing?

Shubhi Rao, Uplevyl (02:57)
Yeah, you know, now sitting where I said I can look back on my life and think everything in like they say happens for a reason. And you know, when I was with Google, I was working on a very large fintech project. And we were really, really struggling with the data sets because the data was messy, it was missing, it was gnarly. and at the same time, I was very fortunate to have sort of a front row seat to

Seeing where our investment dollars were going, and we're investing pretty heavily into AI. Now, of course, everybody talks about it, but AI has been around for a long time. for all of us who use Spotify and Netflix and Google, of course, but it was more, you know, in the background, it wasn't really in the hands of the everyday person. At the same time, being a senior woman leader, very focused on developing, you know, women along the way, making sure that we have great.

succession plans and really deep benches and we're really, you know, working to make sure that our women leaders get opportunities. And so just I think I just happened to be at the right place at the right time where a lot of these things were happening around me. And as I was working on this project, I was thinking to myself, well, if AI is going to be really this big, and in the end, AI is only as good as two things, the data sets that power it, and then the algorithms that you build.

Julia Lach (04:14)
Mm-hmm.

Shubhi Rao, Uplevyl (04:15)
Well gender data sets was the question I had in my mind. Where are the gender data sets? If we're going how is it gonna serve women? Because in essentially to serve women, we need relevant data sets that really understand their our lived experiences. And I think that was the kernel of a thought that eventually now has become Uplevyl.

Julia Lach (04:35)
Very exciting. Bias is often like discussed as if it were only a technical flow. But what I hear you saying is that it also is about like whose reality is actually present in the system. So let's unpack the term center of Uplevyl work, talking about the term gender data.

What does gender data mean in practice and what do people often misunderstand about it?

Shubhi Rao, Uplevyl (05:00)
You know, I often say women are not one homogenous group, right? we can show up as caregivers, as mothers, as employees, as women leaders, as survivors, as investors. And our lens, our frame can be quite different depending not only

Julia Lach (05:05)
Mm-hmm.

Shubhi Rao, Uplevyl (05:18)
you know, relative to men, but even amongst ourselves as women. And so we really need to capture those experiences of when a woman becomes a mother, let's say for example, what does that really mean to her economic wellbeing? How is she now thinking about is she gonna take a break, career break? Is she gonna have to r think about you know, setting aside funds for her child to go to college? You know.

Julia Lach (05:44)
Mm-hmm.

Shubhi Rao, Uplevyl (05:44)
What are some of the decisions she's gonna make across her life that is gonna in the end affect not only just her economic well-being, but many of the decisions and choices she makes? Another one would be around, if you think in and I said this the other day on another show, which was you know, when you look at the great wealth transfer happening, and women, you know, being the recipient of of these large sums of money, but more.

in the latter part of their lives. Well, the reality is that women live longer, but unfortunately we end up spending a lot of money towards the back end of our lives on health care, which then means that we need a very different investment profile. We think about money differently, right? Another example would be just healthcare. Just it's very obvious, just we have just different physiology. And you know, whether it's breast cancer, whether it's menopause,

Julia Lach (06:25)
Mm-hmm. True.

Shubhi Rao, Uplevyl (06:36)
Even heart issues. again, we have to capture those lived experiences not only just across a bucket of women, as I would say, but even you know, women, let's say of my race, brown women may have very different health symptoms and issues relative to other races. So it's really figuring out that first of all, we do need gender data assets.

Julia Lach (06:53)
Mm-hmm.

Shubhi Rao, Uplevyl (06:58)
to address those unique lived experiences women have as a virtue of the different roles they play. But then within that also you need to really think about the segmentation and the cohorts that come about.

Julia Lach (07:09)
Mm-hmm, mm-hmm. Understood,

What would you say, which life events or decisions show most clearly why mainstream AI systems often do not understand women's realities well enough? You mentioned a few examples.

Shubhi Rao, Uplevyl (07:21)
Well, you know, there are so many, but I c I'll go back to just unpacking the example I just gave a bit more on you know the wealth transfer. So in it's not only the fact that you know, with the boomer generation now that retired a few years ago and were in a place where this great wealth transfer is happening from predominantly the boomer men to the women.

Julia Lach (07:30)
Mm-hmm.

Shubhi Rao, Uplevyl (07:46)
And so these women are maybe in their late 60s, 70s, early 80s, having to deal with widowhood. Widowhood simply is not just a transfer of the money, but when you become a widow, and by the way, also a widower, but let's just stick to this example of a widow, she has to handle over 22 plus different things in life, everything from death certificates to estates to probate planning to

social security, right? Like there are so many, many things that need to get handled. All right. So now you're, you know, for most part, maybe your spouse did it. Maybe you don't have a relationship with that financial advisor. You're having to handle all this at this point in your in your life And so this is where you know technology can really help because now you're building a product

Julia Lach (08:23)
Yeah.

Shubhi Rao, Uplevyl (08:34)
through her lens, going, okay, the persona to use this product is a woman maybe in her late 60s and beyond, a widow. You know, maybe this is her understanding of finance. This is all the help she needs in the first 90 days. Then she this is all the help she needs in the first year. Then this is all the help she needs beyond. But you know, there's and by the way, it's not one size fits all because

Julia Lach (08:44)
Mm-hmm.

Shubhi Rao, Uplevyl (08:59)
At least in the United States, so much of it is also very specific to your jurisdiction, specific to your own economic background for what you're eligible for and what you're not eligible for. Where they're just so much and again, AI is a great tool to handle this complexity, but somebody has to go away and build this tool through that persona. Right? So you might so I don't, you know, I that's what I mean. Like when you're thinking about really building gender-focused tools, well, then you have to look at these.

Julia Lach (09:19)
Mm-hmm.

Shubhi Rao, Uplevyl (09:29)
Different personas and the problems that they have and trying to solve. But it's not just that the widow has the problem. The problem then lands actually on the shoulders of the financial advisor who's trying to help this widow. But now the financial advisor is like a quarterback trying to coordinate accountant, the tax planner, the estate planner, the insurance person, right? And at the same time trying to help this widow.

So, what happens is the system starts to break down because you just don't have enough hours in the day, especially in the period we are in right now, where we're going to see an increase in the boomer widows between now and 2045. So, you know, in an environment where you have the demand going up, but at the same time, the other aspect of the financial industry is also.

Julia Lach (09:59)
Yeah.

Mm-hmm, mm-hmm.

Shubhi Rao, Uplevyl (10:17)
That they're going to be losing advisors. We're going to be 110,000 advisors short in the next four to five years. So you have this really interesting, you know, phenomenon happening whereby both the supply is increasing, the demand is increasing, but it's a widow again who'll get impacted if she doesn't get all the services she needs. So recognizing the macro trends, recognizing the needs.

Julia Lach (10:39)
Mm-hmm, mm-hmm.

Shubhi Rao, Uplevyl (10:44)
is how Uplevyl approaches w what problems we decide to solve.

Julia Lach (10:47)
Mm-hmm, mm-hmm. Super exciting. This is just where the topic becomes very human. A model might see the same event, but it may just miss how differently that event plays out for women. Exactly,

Shubhi Rao, Uplevyl (10:58)
Divorce is another one, right? Again, is a it's a very

different outcome oftentimes for the man than it is for the woman.

Julia Lach (11:05)
True, true, And that brings us to the next question of what kind of platform you're actually building. Uplevyl is not only building an AI product for individual women, but a platform for organizations and women's communities. Why did you choose this model? And what becomes possible when organizations bring their communities while Uplevyl provides the intelligence infrastructure?

Shubhi Rao, Uplevyl (11:28)
Well, broadly speaking, you have two models, a B2B model and a B2C model. with the B2C model, I thought about it. You know, what problem set would I try to solve? Because it's like boiling the ocean. That's one. But more importantly, there are incredible organizations out there that are focused on helping whether they're women employees.

The women clients, their women patients, the women survivors. You know, there are thousands and thousands of organizations who already are working with these women. But what we now need in this day and age is to make sure that we equip those organizations with the right tool to address this audience. Because if they just use the generic tools, and by the way, the world of AI is moving super fast. And so

If we don't build those tools now, then we will just default to using the one size fits all tool. And so this is why I thought, no, the the better answer really is because we're technologists, we really, what do we know about helping truly, you know, women patients or women caregivers or women survivors? We don't. We know our lane, and our lane is to build good tech. So how can we then partner with the organizations?

Julia Lach (12:22)
Mm-hmm.

Shubhi Rao, Uplevyl (12:42)
And oftentimes you know, we have organizations like you mentioned about on the community, it's all well, it in in in the sense that it's a community, but it's really them bringing their women together to solve big problems. Right? So there might be a group of women on the platform whereby they are solving their researchers, or they might be policymakers, or they might be advocates, right? So so these are

Women that are all across the country, and sometimes there are men too, not just women, but they're trying to solve a gender issue. Right? So you get both people who are really trying to solve the gender issue, or people who are in the like serving their audience, which you know predominantly could be women, but it could also be men. But we are we build the product through that gender lens. And what I mean by the gender lens really you know, really applying the feminine.

Julia Lach (13:15)
Mm-hmm.

Shubhi Rao, Uplevyl (13:34)
principles around thinking through safety, security, data privacy, how we train the algorithms, depending on the application, maybe they need to be trauma informed or it needs to be you know, much more understanding and empathetic and sympathetic. the training of those models is very different than models that predominantly in the B2 space, I'm not saying always

But are built to extract data to be able to sell it to the advertisers, right? That's a different model. And and in that scenario, you have one big platform, you're trying to bring billions of eyeballs in together so that you can obviously gotta make money, and that money is advertising. Our model is super transparent, work with the organizations that are really trying to help and advance and accelerate women or solve.

the big gnarly problems and not be in in the game of trying to extract and violate data, but to build safe platforms.

Julia Lach (14:30)
Mm-hmm. Super, super exciting. Should we when AI supports women in sensitive areas such as rights, wealth, health, career decisions, or life transitions, what does trust need to look like? And where should the line be between AI supporting a decision and AI making a decision? What's your take on it?

Shubhi Rao, Uplevyl (14:49)
Well, I am very much in the camp. I'm gonna work backwards that AI should help you to make a decision, but it should not make it for you for sure. in the end, I still believe that you do need to trust human experts. You know, especially when it comes to health, you should definitely see the physician. If you're you know talking about, you know, needing help as a caregiver or a survivor or whatever, you should go talk to

lawyer or advocates right so those experts are there and you should use them. let's use an example, we're actually working with a large organization that focuses on gender based violence and one of the projects that we're working with them is helping survivors access

their rights, their employment rights, as it workplace rights, as it relates to paid leave, accommodation, unemployment insurance, and discrimination, right? And these rights are a mess because they're s well, we have 90,000 plus jurisdictions in the United States, but within that these are spread across thousands and thousands and thousands. And each one of those rights has

hundreds of different eligibility criteria. Plus, if you were to click on any other websites, they're all in very strong legalese that the everyday citizen would really struggle to understand. So this is a very good application whereby a survivor could go into the platform, say, hey, look, you know, what are her circumstances be able to understand does she even have the rights? What do those rights look like? What kind of forms would she need?

What kind of a letter? So now she's equipped so that she can then go talk to the advocate in if necessary or to her employer, who whatever whoever she feels comfortable with. But it a it saves the the advocate or the employer time because they have everything. But on the other hand, she also feels pretty confident that she has all the rights that she can exercise and then that keeps her employed.

Julia Lach (16:38)
Yeah.

Mm-hmm, mm-hmm. True.

Shubhi Rao, Uplevyl (16:47)
Right. And and she can

so that makes a huge difference. And so that's but how do you do it? You need to then build tools that allow people to come in to be truly anonymous, to be able to deprecate all that the whole chat history. You know, you're not using that to take back and to retrain your models or you know, make it available commercially to others, right? You have to be really respectful.

and create that psychological safety with this tool. And so there's a lot that goes into just over communicating throughout the platform that you are in a platform where, you know, we're really trying to work very hard to provide you with that safety and security.

Julia Lach (17:26)
Mm-hmm. Exactly. Shubhi, how do you build women centered AI without treating women as a homogeneous group? You mentioned it before.

Shubhi Rao, Uplevyl (17:35)
Yeah, we well from the very beginning, what we did is we wanted to address women's lives professionally, personally, and financially. And since I I think I told you that gender data sets don't just exist. You can't just go purchase them and do anything. So we had to build first party data sets. And what that means is you look at the professional side and use you know, we have subtopics under that.

Julia Lach (17:45)
Mm-hmm.

Shubhi Rao, Uplevyl (18:00)
everything from you know leadership to building confidence to thinking about promotions, you know, having difficult conversations, right? Like all the topics that you would, you know, getting on boards. But the whole idea was to build the foundational LLM large language model that was maybe a inch deep and a mile wide. And and how so that we could really address all the main topics that impact women professionally, personally and financially.

And how did we do that? We brought in contributors, like truly women experts. They were physicians, lawyers, psychologists, board leaders, wealth advisors, like truly women who have domain expertise to be able to provide their knowledge base to build this foundational layer. Sixty-seven percent of the contributors we brought in are are of diverse backgrounds.

Julia Lach (18:51)
Mm-hmm.

Shubhi Rao, Uplevyl (18:51)
And so

the foundation of Genie reflects those not only the different expertise and the and and the experiences, but also the fact that first of all those people came from across the nation and some from the United Kingdom and parts of Europe, but also bring in very different intellectual, cultural, racial frames into and into the diversity of that.

Genie has today.

Julia Lach (19:17)
Mm-hmm. Yeah, super, super interesting. Women centered AI has to hold complexity, has to hold culture, life stage, different realities. exactly. And this just leads to the bigger question you wanted to bring into this conversation. In our first conversation, Shubhi, you said something that I really wanted to bring into this episode, which is women cannot only criticize the bias in AI.

Shubhi Rao, Uplevyl (19:24)
Yeah. That's right.

Julia Lach (19:41)
We also need to build, lead products, and use our voices inside organizations. Why is that distinction so important to you?

Shubhi Rao, Uplevyl (19:49)
Look, if you l look at where the markets are today, billions and billions and billions of dollars are going into AI. I think it's unprecedented to really just see how fast AI is moving. And I'm not discounting at all the fact that, you know, the issues being raised are a hundred percent legitimate.

Right, okay, so you raise the issues, but raising them is a really good first step. But what do you do with it? the answer might be let's get regulators to step in and help. But how does that really work? Because we can't even get anybody to agree on, you know, nuclear weapons across the globe. So when you think about ev the whole globe agreeing to

Some sort of a singular mandate around how AI gets regulated. I don't know if that's gonna happen in my lifetime. But let's work a little bit closer to home. In the US alone, I'm not sure whether we'll just get one big federal mandate or statute or law, because again, like I told you, we have 90,000 plus jurisdictions and everyone may come up with their own, right? And in some ways, as a capitalist society.

Julia Lach (20:51)
Yeah.

Shubhi Rao, Uplevyl (20:54)
I think we're gonna resist it. Even me as a technologist, I don't wanna be burdened with having to comply with thousands of laws across the US without I I'm happy to comply, but give me one solid law across and great. I can build a platform that fits, but if I have to build a platform that now fits 50 different statutes, it makes it really expensive and difficult and painful to do so, right?

Julia Lach (21:19)
Hmm.

Shubhi Rao, Uplevyl (21:20)
And so

Julia Lach (21:20)
Yeah, yeah.

Shubhi Rao, Uplevyl (21:21)
there are real practical limitations. And then the other thing, by the way, is that this is the issue if the US does it, but China doesn't do it. Right? So there's a bigger issue from a global perspective, which is well, then are we disadvantaging the technologists in the US and then China may win the race, right? So so there are all these these are like all the different tensions that we're dealing with right now because laws unfortunately don't permeate across the globe.

But AI permeates, right? Like it doesn't know boundaries, it doesn't know, it it can just goes around any everywhere. So so I think that there's that big disconnect. So having said that, I do agree that we absolutely have to raise the issues, but we also have to be part of the solution, you know? And we know some of the issues around the solution. So where I go back, and maybe I'm being naive, but as a technologist, is that

Julia Lach (22:03)
Mm-hmm.

Shubhi Rao, Uplevyl (22:09)
Why not just build it right in the first place? Right? Like how do we build it such that we can still be super competitive around the globe? We can we can use our own common sense and our own moral compass and and and recognize what's right for children. How do we make sure that we're respecting you know our user base? How are we protecting the right? Like I mean, that I don't think really is something that

Julia Lach (22:17)
Yeah.

Shubhi Rao, Uplevyl (22:35)
necessarily has to be regulated, it should just be what good companies should do. and how do you do that? Well you need people we all don't have to go and be the builders of technology, but we sure can use our voices and make sure that we get included and have a seat at the table as the product is being designed, as processes are being, you know, reengineered.

as data is being collated and curated, you know, as algorithms are being built, right? That is where we should really use our voices even stronger to ask how is this being done? How was this being tested? How was this sourced? You know, who is this being tested on? You know, like all that I think is how we really need to insert ourselves. and just be much more, I mean aggressive seems like a pretty strong word, but we really probably need to

Julia Lach (23:20)
Mm-hmm.

Shubhi Rao, Uplevyl (23:25)
really ensure that we don't shy away from taking those steps.

Julia Lach (23:29)
Mm-hmm, mm-hmm. More dominant. Yeah, definitely. This just feels very close to the heart of she builds with AI. We can and should name bias, but we also need women in product, in data, in strategy, in governance and leadership roles. So let's turn this into something practical for the women listening.

For women inside companies, communities or startups today, where do you see the greatest leverage to shape AI before it becomes embedded in products, workflows or decisions? What questions should they ask? What rooms should they try to be in? And what kind of ownership should they claim?

Shubhi Rao, Uplevyl (24:07)
Yes. I don't think there is a single function within corporates today where we are not seeing a complete process reengineering happening. And I think anybody and everybody could participate in it. And so as we're going through this again, unprecedented process re-engineering, there's a there's definitely a place for us to ask the question.

Julia Lach (24:20)
Mm-hmm.

Shubhi Rao, Uplevyl (24:28)
Well, this is how processes were, legacy processes. With technology, this is what maybe the new model is going to look like. But I think we need to play various scenarios in it. And we should force that question to say, well, this is one way to do it. Because maybe some big consulting firm can walked in here and you know said this is what we should do. But we should say, Well, we should look at the different alternatives and pressure test it to make sure.

it really works for your whoever you're serving, right? And it is very inclusive and it's thoughtful. and and you know it's not it's been really designed with the long term, with a longer term perspective, not just short-term profits. Right. So so I think that's clearly one low-hanging fruit because that's just what's happening across the globe as people are really starting to think about where do I use human capital?

Julia Lach (25:16)
Mm-hmm.

Shubhi Rao, Uplevyl (25:21)
And where do I use technology? Right. That's that's a conversation happening in every boardroom, in every C suite, every leadership meeting, right? Every team meeting. That's across the board. So I think that's one low hanging fruit for sure. The other place, of course, is you know, when I think about folks that are on the board, right? That's let's start at the top and then we'll also work into the companies. But I think today it's

Julia Lach (25:31)
Mm-hmm.

Shubhi Rao, Uplevyl (25:46)
It's shifted so dramatically because now boards have to think about this from a risk lens, right? And the risk is is on many levels, not just financial risk, there's a legal risk, there's reputational risk. So really understanding right what is the governance, because that is a job of the board, is governance. And so how do you think about now this new model of governing when when you when you look at the companies today, again, like I said.

You could have human capital, you could have digital employees. And how are you, even that is an example, how are you starting to think about it? So it's really going back now and rethinking what is how what does that organization model look like? What as a result, what kind of risk, how you know, what functions are being used? we are using technology? As a result, what does that how does that shift our

risk parameters, reestablish what your risk appetite is now. Right. So so that whole conversation about risk that we've had for a long time that yeah shifted a bit, but not so much. I think now this puts the whole risk conversation in a whole different light. So I think that's another place where you know as women either could be board members or you're presenting to the board, right? You have the opportunity to present to the board, then you have the opportunity to

be able to present your work through those different lenses and take advantage of those opportunities to do so. And then, you know, obviously as you're a senior leader or you're in, or you know, you're somebody who's an individual contributor, it's continually asking those questions around, and those questions, by the way, don't dramatically change from one function to the other. Because where is the data set coming from this?

Julia Lach (27:29)
Mm-hmm, mm-hmm.

Shubhi Rao, Uplevyl (27:30)
How are these

algorithms being trained? You know, who who's testing this? Right? Like these are all the fundamental questions that it doesn't matter whether you're in HR, you're in marketing, you're you know, in product development, it's all the same. So so that but all this can only happen if how you're thinking about upskilling yourself, like because what we're seeing now is that all the jobs that are really washrooms, repeat jobs, if you will.

Julia Lach (27:35)
Mm-hmm.

Mm-hmm. True.

Shubhi Rao, Uplevyl (27:57)
We're really easy to get technology to do it. So the higher cognitive order jobs I still believe humans will have. And so we have to just make sure that we are now upskilling ourselves so that we can fully understand what this new world that is very technology enabled looks like.

Julia Lach (28:06)
Yeah.

Mm-hmm. Mm-hmm. Yeah, definitely agree. I just like this because it makes this call to build very practical. Because building AI is not only writing code, it is also defining the problem, choosing the data, asking where the data comes from, shaping the product, asking who benefits, and just noticing who might be left out. And that brings us to the

Shubhi Rao, Uplevyl (28:37)
Yes.

Julia Lach (28:39)
Bigger future you are working toward. Five years from now, what would make you say women did not just adopt AI faster, we actually helped build it better.

Shubhi Rao, Uplevyl (28:49)
Look, I think what I really want is AI and its applications to be accessible. I'm worried a bit today that at the pace AI is moving, that you know women don't fall behind. and let me give you some examples. When you

Julia Lach (29:07)
Mm-hmm.

Shubhi Rao, Uplevyl (29:10)
I if you just think about just simple things where if you look at Claude, Claude is a great product that you know enterprises are using pretty significantly, you know, to develop content, to work with Excel, you know, to help with your productivity. But each day, Claude is getting more and more sophisticated. And you can really start, you really need to understand it's a whole new discipline in a way. It really is a whole new discipline.

So in a few years ago, it's like, I know how to write a prompt. Great. But now nobody cares about that. It's a given. Of course, you have to know how to do Ruddha. But it's now really being able to use all the various capabilities and features that Claude has in a very smart way. So you don't necessarily have to be a coder, but you can get Claude to write the code for you to enable whatever application that you're doing, that historically you would have had to wait for the next new release of.

Julia Lach (29:37)
Mm-hmm.

Shubhi Rao, Uplevyl (30:04)
SAP or workday, you can actually build that now within you know the application that you're using. so whether it's co-work or it's using Claude code and understanding, you know, how projects work, how archetypes work, all of those things now are going to be part of your vocabulary and you better get pretty fluent at it, right? Like that's just one example. So and that's what is gonna make all the other tools accessible to you because the gating factor might be.

Julia Lach (30:06)
Mm-hmm.

Mm-hmm, mm-hmm.

Shubhi Rao, Uplevyl (30:31)
your knowledge base around Claude. Another different example, which I just gave you, was was earlier working with this this large organization to help survivors. All of a sudden with technology, we're making those workplace rights accessible to the everyday person. Now that person can come in, understand their rights, and as a result be able to

Julia Lach (30:51)
Mm-hmm.

Shubhi Rao, Uplevyl (30:56)
keep their job or be able to, you know, protect themselves or be able to take action that they otherwise couldn't have. Right. So so that to me, and and I and I and and this is what I really want, right, right, is more and more women to really think about and men by the way, how how do we use these tools to make products, services, data that was inaccessible?

Julia Lach (30:58)
Yeah.

Mm-hmm. Yeah.

Shubhi Rao, Uplevyl (31:20)
Because it was fragmented, it's messy, it's everywhere, being able to actually create because we have these foundational LLMs, great, great data in them, but they're very generic. What we now need is very specific, curated, collated data sets to address the needs of those various populations. And so that to me, I think, is a huge opportunity set for us to think about how we can use technology in a smart way.

Julia Lach (31:37)
Mm-hmm.

Shubhi Rao, Uplevyl (31:46)
to build a way build the tools to make it more accessible to the to the everyday person so that the gap doesn't become so huge where the haves and the have nots where where people who are really in the whole AI space and people who are out of the AI space. That gap will just widen otherwise.

Julia Lach (32:02)
Mm-hmm. Yeah, This is just a beautiful place to end, Shubhi, because it just brings us back to possibility. AI is not only a technology we adopt, it is something we're shaping. The data we value, the products we build, and the voices we just bring into the room. That was my conversation with Shubhi Rao. Thank you so much, Shubhi.

What I loved about this conversation is that Shubhi brings AI back to something very human, the data of our lives. Not only the data that is easy to collect, but the experiences, decisions, transitions, and risks that often remain invisible when systems are built around a default user. Her work with Uplevyl is a reminder that AI is not only neutral simply because it is technical.

It carries the assumptions, gaps, and power structures of the data and systems behind it. And if women's lived experiences are missing, AI can become faster and smarter while still failing to understand half of the world properly. But what stayed with me the most is Shubi's call to agency. Yes, we need to question bias and we need governance, privacy, and accountability, but we also need women who build.

Women who lead products, women who ask better questions women who bring lived experience, domain knowledge, and courage into the rooms where AI decisions are made. And that is exactly why SheBuilds with AI exists, to spotlight the women who are not just reacting to technology and its change, but actively shaping it. And to remind all of us that building with AI does not always mean writing the model yourself.

It can mean defining the problem, designing the product, questioning the data, protecting the user, and making sure the future includes more of us from the very beginning. If this episode stayed with you, shared with one woman, founder, leader, operator, product person, or colleague who needs to hear that she has a role to play in AI now, not someday.

And if you enjoyed this conversation, follow SheBuilds with AI, leave a rating or review, and help more women discover these stories. Until next time, stay visible, build boldly, and help shape technology in a way that works for all of us.