22. Data, AI & Business Outcomes - A Talk With Amit Shivpuja (Walmart) & Madhavi Rajan (Rackspace Technology)

Listen and get insights into "Data, AI & Business Outcomes" in this talk with Amit Shivpuja and Madhavi Rajan. Amit is the Director of Data & AI Enablement at Walmart, and author of the book: The Data & AI Compass. Madhavi is the Head of Product Strategy, Research & Operations at Rackspace Technology, and she is a Member of the Board of Directors at Women in Big Data.

In this episode you will gain insights into the intersection of AI, Data, and Business outcomes, emphasizing the importance of AI and data literacy in organizations. You will learn that many enterprises, especially within the Fortune 100, are still in the early stages of AI adoption, shifting from hype to practical execution. The conversation highlights the need for categorizing AI initiatives to improve operational efficiency, customer engagement, and product strategies. Central to the dialogue is the emphasis on strong leadership and organizational alignment to effectively navigate the continuously changing AI landscape and drive meaningful business transformation.

"I would say AI isn't a tech in initiator. It's actually a leadership shift. So, we are not just like rolling out the models or the product; it's actually much more than that. It's is values, it's risk, and it's trust. And that takes not just innovation of your product but it takes like alignment, and then the sponsorship and executive courage." - Madhavi Rajan.

Guest Info

• LinkedIn: Amit Shivpuja
• LinkedIn: Madhavi Rajan

Resources

• Book: The Data & AI Compass (Amit Shivpuja)
• Website: https://daicompass.com
• Book: The Thinking Machine (Stephen Witt)

Chapters

• 0:00 Intro
• 1:14 Current State of AI & Where Are We Going?
• 4:30 Madhavi's role in shaping the AI strategy at Rackspace Technology
• 6:49 Amit's role in shaping the AI strategy at Walmart
• 8:01 Amit, author of the book: The Data and AI Compass
• 8:30 Key infrastructure and data dependencies before you can start leveraging AI
• 10:16 Book: The Thinking Machine - Stephen Witt (The story of Nvidia)
• 11:03 AI, Business Leaders, and Professionals
• 11:47 AI literacy
• 12:39 Is AI going to take my job?
• 13:22 Examples of AI projects with succesfull business outcomes
• 16:11 Who is a women in big data that you admire and why?
• 17:09 What is the best career advice you have ever received?
• 20:37 Outro

Transcript

[00:00:00] Intro: Hello. Welcome to the Women in Big Data Podcast where we talk about big data analytics and career topics. We do this to connect, engage, grow, and champion the success of women in big data. 

"I would say AI isn't a tech in initiator. It's actually a leadership shift. So, we are not just like rolling out the models or the product; it's actually much more than that. It's is values, it's risk, and it's trust. And that takes not just innovation of your product but it takes like alignment, and then the sponsorship and executive courage." - Madhavi Rajan

In this episode, we talk with Amit Shivpuja and Madhavi Rajan about Data, AI and Business Outcomes. Amit is the Director of Data and AI Enablement at Walmart, and author of the book: The Data and AI Compass. Madhavi is the Head of Product Strategy, Research and Operations at Rackspace Technology, and she's a member of the Board of Directors at Women in Big Data.

Let's start.

[00:01:08] Desiree: Amit, Madhavi, welcome to the session about Data, AI and Business Outcomes. What is the current state of AI and where are we going? 

[00:01:17] Amit: Thank you, first of all, for the opportunity. 

[00:01:19] Desiree: You're welcome. 

[00:01:20] Amit: So, it's definitely a hot societal debate that's going around AI and its impact, the global competition that's happening between countries like US and China.

Even the polarization of opinions that are out there. There's a lot going on. If we look at the globe, if we bring that down, we see that there is still some hype. But organizationally, people are having a lot of conversations around execution, right? How do we make AI available at an enterprise scale? Now, that also means that there's a lot of conversations going around budgetary things like what should the AI budget be? What is its impact on P&L? 

There's also interesting discussion parallel to this by policy makers on enforcement. What kind of transparency should be there? What kind of accountability should be enforced on people? But I think the largest conversation, or the most recent conversation, on AI is around talent and the workforce. Right? What's gonna get automated? Is it just gonna make things more efficient, or are some of the job cuts people are talking about the major impact? But relevant for me being in data and AI governance, is very simply: what's the landscape going to be? What are we going to govern? What kind of roles are we going navigate and enable for the same?

There's technologically still a lot of evolution going on, a lot of new capabilities, and a lot of new things are expected or coming out as we speak. And this has implications on every role I would say. Where my opinion is that people should be both data literate and AI literate to be able to go forward. 

I'll leave it with that as an overall, what's the current state and where things are going. 

[00:03:09] Desiree: Madhavi, something you would like to add?

[00:03:12] Madhavi: Amit, touch the global impact of what's happening with the AI, narrowed it down all the way to the data. From the place where I am, working for a private cloud business "Rackspace". The way I see it, we are really early in terms of the AI adoption. So, living in the Silicon Valley and working with customers who are in sovereign and then oil and gas, healthcare, finance, and all of that, I am in conversations where people are talking about the latest and the greatest of the ChatGPT. I get to see the like a huge variance in the AI literacy, AI adoption and so on. And from that angle, I see we are like at the very early phase in terms of enterprise AI adoption. Maybe 10% or so in enterprises, maybe mostly in the Fortune 100. And a little spilling over to the Fortune 500 where they have the budget to go and spend and build. The rest of the 90% are yet to catch-up. 

It's gonna move fast having said that, because in the past couple of years I've seen people just beginning to talk about AI to more POCs, and now it's getting more into the production phase, right? So it is moving quite fast, but still a pretty early stages. So that's my view. 

[00:04:28] Desiree: Okay. And what is specifically your role in shaping the AI strategy at Rackspace? And what decisions so far have had the biggest impact? 

[00:04:37] Madhavi: Currently, I'm looking at how do we position AI for Rackspace. And the way I see it, and when I stepped in: it was like more scattered. And I see that reflect in many of the customers and other enterprises that I talk to as well. So, then the key thing at the fundamental level needs to be like categorization of how I'm consuming AI. So, it's AI for myself, AI for my customer, and then AI for my product. 

So, AI for myself is like as an enterprise or a medium or a small business: how am I going to improve the operational efficiency of my business using AI? So, that then sits under IT. It's kind of getting siloed and the silos need to be broken down. Like, it's not functional level organization anymore but then AI for business outcomes, right? So, then AI for myself is there. 

And then AI for my customers is: how I'm interacting with my customers? How do I personalize my customer experience? Like if I'm a finance segment and I have like folks who have accounts with my bank, my investments and my relationship with the bank is very different. And if I'm gonna be a VC firm, my relationship with the bank is very different. So, then whatever portal a bank offers would not apply for customer A versus customer B. So, how do you go customize it? And AI is just going to accelerate that experience. So that's me and the customer AI.

And then AI as a product: like right now, a lot of the AI products are based on either chatbots or co-pilots. So, it's Salesforce including agents in their products, or Microsoft including co-pilots in their products. But then a lot of AI products and how people have to translate into revenue generating model is yet to come. And that's where the 10% sits right now.

So, my role right now is to take the promise of AI into more of an executable business strategy. Have these separations out, and how do I make sure I take AI to convert into business outcomes or revenue generation for the business? So it's been a very interesting learning experience for me. 

[00:06:49] Desiree: Amit, what is your role in shaping, AI strategy at the company you are currently working?

[00:06:54] Amit: So, I come from a data perspective, right? I've been leading and being part of data organizations for most of my career. And what I'm currently focused on is given how foundational data is for AI and operations for the organization. In some ways you can say I'm trying to take advantage of the buzz and hype around AI to make sure that the intersection we have of data governance and AI is effectively in place.

We wanna be able to have trusted data that organizations can rely on. You wanna do it in a form that's there long term. That's where the governance and strategy comes in. And, you wanna make it in a form that's AI adoptable or AI enabling. So, a lot of my time and my conversations goes on: how do we bring those three things together, and what's the overlap and how do you do it in a form that doesn't slow down innovation on the organizational side or the operations that need to happen? As well as balance out all the steps you need to take procedurally in order to make sure the organization can continue to do what it needs to do. 

[00:08:00] Desiree: Amit, I saw that you even wrote a book about it: The Data & AI Compass. 

[00:08:05] Amit: So, I was having a lot of conversation with people independent of their technical level because these are things that matter to them. Data matters to them. How do they do it long term strategically with structure, which is governance and the whole AI piece effect? I'll make an attempt of putting some of this in place. So, the book is an attempt for me to give people a compass to navigate the current evolving ecosystem. 

[00:08:28] Desiree: Thanks for that.

[00:08:30] Madhavi: Amit was already talking about the fundamentals. So, what are the key infrastructure and data dependencies before I can start leveraging AI for my business? 

So for me, the real dependency is the alignment between the technical debt that people have in the fragmented data governance, and then the inconsistent access policies between different businesses within a larger organization.

And largely people are wondering and worried about: infra is so expensive and I need racks and tracks of GPUs. And some of the conversations are really basic, like, get folks literate about what AI is. All you need is maybe two to four GPUs at the max. And when you bring it down to that level, I don't mean millions and millions of dollars, it's probably less than a 100k investment to go and build something that will translate into business outcomes for you.

So, education literacy needs to happen in terms of the infra space as well. How people are consuming that infra? The folks who are running like a small business or medium or large enterprises are more looking at: how do I enhance my product? And they don't go to the level of comparing multiple different infra. And so, how do you translate that and get to the language of; if I'm going to create a website and then automation of that website using AI, what is that processing power or the infra requirement that is needed? And can I make that happen with what kind of architecture? It needs to be as simple as that. Sometimes you don't even need the latest and the greatest of infra that is needed.

[00:10:08] Desiree: It's about the fundamentals. And then of course, what is your need? And then you have a use case and you can start with it. 

[00:10:15] Madhavi: Yeah. 

[00:10:16] Desiree: And for the listeners who want to know more about AI, if you really want to start with the basics, I think you should read a book: The Thinking Machine. It's about NVIDIA and it really helped me to better understand why AI is possible and why it is accelerating.

[00:10:32] Madhavi: I've consumed NVIDIA products. And I've been like competing with NVIDIA products, and now I consume more of Nvidia products. So, I have seen the way the transformation has happened. You know, like people say that it's an overnight success or something, but no, Jensen has been building this for 30 years, right. 

And as part of my strategy work, I do quite read about it and the way the market and the industry is going, Nvidia is uniquely positioned. I'm quite excited on what's out there for us the next five to 10 years. 

[00:11:02] Desiree: Amit, what do you see with business leaders and professionals? 

[00:11:06] Amit: I think most large organizations, especially the Fortune, and the public ones, are definitely talking about AI and their organization's involvement with it.

What I've found is that especially the recent generation of AI, you know, GenAI, vibe coding, that at least the initial adoption is pretty quick, right? Somebody can log onto a website and start playing with it. So that is happening on all levels of the organization. I think where it is getting interesting is around the fact of what does that translate into how it'll be adopted and used within organizations? AI literacy is, to me, non-negotiable at this stage. 

[00:11:47] Desiree: And within your company, is there a kind of program related to AI literacy? 

[00:11:52] Amit: So, that's what we are working on right now is, along with giving access to people, how do we educate them?

What's been helpful in this while we create our own, you know, relevant content is the amount of publicly available stuff that's there today. There's a lot of YouTube videos, a lot of Coursera or learning tools. So, I think that's kind of bridging the gap for now. I

But I think the bigger thing is navigating the cultural chain. So that's, I think, what organizations, and us as leaders, are trying to figure out is: how do you nudge the culture, or modify the culture, or enhance the culture so that people see the value and see the need to become AI literate beyond just the hype and the buzzwords? 

[00:12:37] Desiree: Okay, I understand. 

[00:12:39] Madhavi: So, one of the common ones that I see is like: is AI going to take my job?

It is gonna take away some of the jobs, but it is gonna create or rebalance things. It's a tool, it's a enabling tool to make life better. It's not gonna replace you. But it is going to make you more productive, so there is a trust element that needs to be built. You start infusing AI and you see how beneficial or enabling it is. Then it changes slowly and people are willing to adopt it more and more. You have to make the employees feel that you are part of the transformation. And when you say that, then the adoption is much faster. 

[00:13:22] Desiree: Madhavi, can you give an example of an AI project where the business outcome was successful? 

[00:13:28] Madhavi: So, this is more of like AI as a product example from the different categories I mentioned.

One of the interesting part with the customer segment that I work with is when and where do I adopt AI? Then the key thing what we did was: okay, we need to give people something to play with before they go and adopt it. So, what do we do? So, we created like small demos of applications, right? If I wanna do like just a text space RAG, let me just build a text space RAG for you. And then let you ask me questions or let me go type it up.

It's very interesting. I did that in our sales kickoff, training a bunch of salespeople across the organization. And so for them to see and believe and show like it's very simple, it's not that complicated. Okay, let me just take a PDF file, and I've already loaded it, and I've already made it connect to the open source LMS. Now you ask a question, you go type a question. And then it goes and parses through and then replies back. 

One of the common things that salespeople do is file their billing expenses. So, you have this application right here, you just imported a picture file and you go and ask: hey, how much did I pay for that burger? You just ask that question. You don't have to like manually go see it. So, it's as simple as that. And when they could see and sense it, they're like: okay, now I can go and converse with my customer on how it can enable them. So then you create a business model around it, right? 

[00:14:55] Desiree: It's a really interesting approach. 

And Amit, I read, I think it was this week or last week, that now I can order at Walmart via ChatGPT.

[00:15:06] Amit: If you look at today, the entire search as we understood it, has changed. There was a time you would go to to a google.com, you would then search, it would give you links. You would click on the link or multiple links, and you get the info that you want. But now it's becoming a zero click kind of a transaction where you ask the question, you get the info using ChatGPT.

So, what the organization sees is if this is the interface of choice: how do we become a part of that ecosystem? So, the idea was basically that if someone tomorrow says: hey, I'm planning for my kids 10 year birthday party, gimme a list of what items I should buy. They're expressing intent. They're saying: hey, I'm gonna buy these things.

So, that was the forethought the Walmart leadership had. Okay, fine, let's form that partnership with OpenAI so that when somebody says "I want that list of items for my kids" birthday', the actual list of items is there for them to click on and purchase or evaluate. Because, you already know the intent behind the transaction as this becomes the more preferred way of interaction.

[00:16:09] Desiree: It's a great initiative. 

So, I also have some lightning questions. Who's a woman in Big Data and AI that you admire and why? 

Madhavi, I'll go with you. 

[00:16:18] Madhavi: So, for me it's Fei-Fei Lee. She is one of the very few who bridges the academic side of the world. She's like early person in computer vision, and she connects them into the ethics and then the real world influence.

She's built the AI4ALL, where she is ensuring that the inclusion happens. So, AI4All, it's actually an organization which enables students who are in their high school and have gotten their undergrad and they get to build AI experiments, hands-on stuff. So it's measurable outcome that she is producing and she's just amazing.

[00:17:01] Amit: She does an amazing job. So I would echo a lot of what Madhavi said as being that leader or that person I look up to

[00:17:08] Desiree: Amit, what is the best career advice you have ever received? 

[00:17:12] Amit: I think it's a combination of being curious, of building a good support network. And for both of us, Stanford LEAD experience has gone a long way. Lean into that network and give back to that network a lot. So, that's a very powerful ecosystem for our growth. 

[00:17:29] Desiree: I understand. 

And then for you, Madhavi, what is the best career advice you have ever received? 

[00:17:34] Madhavi: Don't just manage your time, manage your influence. 

I think it's also a learning part of the startup, like you have such a limited time, so it's not about packing your calendar or packing your time with conversations or going from meeting after the other. So, it changed the way how I approached any discussion for that matter. And as you grow in your leadership and your career, it's not about the volume, but it's about like one discussion that you are in. What is the strategic presence that you have? How you have been able to influence the direction or being able to provide the right direction to people.

Like most of the times people get so busy with doing the what and the how, they forget to ask the question of: Why are you doing this? When I moved from the startup to Intel, a lot of my conversation used to be why? And people used to get puzzled: nobody asked us that. Yeah, because you folks have been working here for 15, 20 years. You have done certain things in a certain way, but you have to really pass and ask: is it bringing any value to what I'm doing and my end goal? If not, let's just not do it. 

[00:18:44] Desiree: I can recognize. Thanks for that. 

So, we are already coming to the end of the podcast, but Amit, is there anything else that I haven't asked you yet?

[00:18:52] Amit: I think we're at a very exciting time. I know uncertainty and changes can be uncomfortable. But I think for many of us, we are at a very exciting time where not only can we figure things out for the future, but we can actually get our hands dirty doing some of that stuff and see what it means. So, what I'd say is give it a try. And then make your decision on which way you wanna go, I think will go a long way. Give it a shot, it might just broaden your horizons in ways you don't anticipate.

[00:19:25] Desiree: So, just try! 

[00:19:26] Amit: Yes. 

[00:19:27] Desiree: Okay. 

And for you Madhavi, is there anything else you want to share with the listeners? 

[00:19:32] Madhavi: I would say AI isn't a tech initiative, it's actually a leadership shift. So we are not just like rolling out the models or the product, it is actually much more than that: it is values, it is risk, and it is trust. And that takes not just innovation of your product, but it takes like alignment, and then this sponsorship, and executive courage. 

A lot of AI initiatives are stuck in the POC stage because it's not getting the buy-in from the executives. So my ask is for the listener to your podcast, if you are leading in this space, your job is not just to launch the tech but it is to lead the transformation. So, you need to have that conscious mind shift. And when you do that, AI and algorithms, LLMs, SMLs, and everything, it's just a small box that gets rolled under it. 

[00:20:22] Desiree: Well, thanks for that. 

So, Madhavi, Amit, thank you very much for joining this podcast. And, I hope that sometime in the future we can talk again. So thank you very much. 

[00:20:32] Amit: Thank you. 

[00:20:33] Madhavi: Definitely.

[00:20:37] Outro: Thanks for listening to the Women In Big Data Podcast. For more information and episodes, subscribe to the show or contact us via datawomen@protonmail.com

Tune in next time!