Leveraging AI to Enhance Diversity in Business Leadership

Malur Narayan Transcript

Intro - 00: 00:01: Welcome to Beyond The Screen: An IONOS Podcast, where we share insights and tips to help you scale your business's online presence. Hosting genuine conversations with the best in the web and IT industry and exploring how the IONOS brand can help professionals and customers with their hosting and cloud issues. I'm your host, Joe Nash.

Joe - 00: 00:22: Welcome to another episode of Beyond The Screen: An IONOS Podcast. Joining us today is Malur Narayan, MBA, a technology leader who champions diversity and equal opportunities in business leadership. He's the Co-Founder of Strides AI, a platform that connects leaders from underrepresented minorities with senior management teams, with the goal of improving the diversity of senior leadership in business through the help of AI. Malur is also VP for Sustainability and Emerging Tech, GTM at Tata Consultancy Services, TCS, leading the emerging tech market and solutions across North America. When he's not in the office, Malur is on the Board of Directors for a number of companies also supporting underrepresented communities, be it showcasing community sports teams in India or supporting minority college students aspiring to leadership roles in Texas. Malur, thank you so much for joining us today. Welcome to the show.

Malur - 00: 01:06: Thanks, Joe. Really glad to be here. Excited to be here.

Joe - 00: 01:09: We like to start our guest with a question about their professional histories, their careers. Looking at yours, one thing that always immediately jumps out is, you have worked all over the world from places like Ottawa and Brazil to Mumbai to Texas. So I want to start by asking, with your experience in all these countries, all these different working environments, how has working across these various regions and locations affected the way that you perceive the impact of your work?

Malur - 00: 01:33: Yeah, that's a great question. I think when I started out, I obviously was very naive and, you know, I had a very single worldview of how things work. And one of the great things about working and living in many different countries is that you get to experience firsthand the social life of different stratas of society and how business and work impacts society in general. What I found was at the most basic level, every culture, every country is trying to do the same thing, is improve their lives, improve the lives of others, of their families, of their friends, and grow together. And in terms of work itself, I found, you know, obviously, when you're living in the US, working in the US or Canada, it's a very individualistic culture. The focus is more on individuals. So there's a lot of effort put on community and teamwork, et cetera, on the outside to bring that out. Whereas in Asian cultures or even Latin American cultures, there's greater emphasis of working together and the focus on individuals is less. So, you know, there's some differences, but the way you work with the customers, the way you work with businesses is also different. The way negotiations are done are very different between those. So it's been a great learning experience.

Joe - 00: 02:45: It's a fascinating point about the difference between individual and community cultures in those different regions. That's really interesting. So across your career, you've been at Tata for, is it 13 years now? 13 years?

Malur – 00: 02:56: Yeah. Awesome.

Joe – 00: 02:57: So obviously that's a long time in the tech industry and associated spheres. What first attracted you to working in the tech industry?

Malur - 00: 03:03: I did computer engineering way back. And when I moved to Canada to do my master's in AI and Machine Learning, this is way back in 1991, when it wasn't so cool. Somebody once told me that, you know, AI is a bad word back in those days. So don't use it if you want to get employed.

Joe - 00: 03:21: Absolutely. It's actually funny. A thing I tell current students quite a lot is that even when I studied computer science, which was not that long ago, it was in 2013, we were told by our AI instructor at the time that neural networks were a dead end and they're not going anywhere. And I always think the various cycles of AI winter, AI boom are fascinating, especially in this current moment. So don't you say it was 1996?

Malur - 00: 03:43: 1991, 1991, yeah.

Joe - 00: 03:45: Incredible.

Malur - 00: 03:45: So, of course, I dropped out of a PhD in AI and decided to go work for a company called Nortel or Bell-Northern Research, the Canadian version of Bell. And that was an interesting experience because I got to use a lot of my AI and ML skills to reverse engineer software. And I've always been a techie. Technology has always fascinated me. And even when I was doing other roles, business-oriented roles or sales roles or business development, tech has always been at the forefront.

Joe - 00: 04:10: Yeah. And I guess going to that experience where you had to put it on the back burner for a bit. These last couple of years, or I guess... Coming on to five, 10 years, but especially this current LLM moment must be very exciting for you and bring back some real memories.

Malur- 00: 04:22: It is. And LLM is exciting because it's bringing AI to the mainstream. It's finally giving people an idea of the capability that AI can have. And the impact that it can have in terms of productivity improvements, in terms of helping get away from doing the mundane tasks and elevate themselves to more value-added tasks, and let AI do all the grunt work.

Joe - 00: 04:41: Yeah, for sure. You mentioned the impact of AI there, so that's a very good opportunity to dive into Strides. So in the intro, we heard that Strides is a platform that you're creating to connect leaders, nonprofits, and communities to senior management teams to address diversity issues, and that's AI-assisted. Can you tell us a little bit about, as a co-founder, someone starting this business, what were your aims when you started Strides AI?

Malur - 00: 05:01: My co-founder, Rika Nakazawa, and I, when we started this, it was really focused around impact. The idea was to see how we can use technology to help solve real-world problems, such as diversity on boards. So that's how it started. It started with a focus on gender diversity, and it's even today that happens. A lot of those roles, board-level roles, or very senior C-suite roles, are filled by people you're comfortable with, people you trust, people you know. So really, it was a relationship thing. Obviously, you have to have the skills, etc. So what we looked at is, how do things happen in the real world? In the old days, or even to a large extent even today, those relationships are built on the golf course, in country clubs, sailing grounds, in social settings, right? And then once you get comfortable with the person, people hire the people that they trust, and therefore, invariably, 80% of those roles are filled by people you know. And we looked at technology as, how can we build a platform that enables people to build relationships without necessarily having something hanging over your head that says, “hey, you've got to hire that person”. It's more about, how do you build relationships at the executive level? And what we found was many of the board members and C-suite leaders, they were keen to connect with the next generation of leaders. But because of their limited time, they wanted to make sure that they maximize that time effectively, that they talk to the most relevant people, people that are most interested in that particular role. So the platform does three things. One is once you create a profile, it recommends what types of people you should connect with. It seamlessly enables you to set up meetings and have video conversations with specific individuals once they accept. And then instead of trying to meet as many people as possible, the focus is on meeting the same person multiple times so that you build a more long-term relationship as a sponsor, as a mentor or whatever. So that's the idea. And we have almost a thousand people on the platform, over 67% of women, over 40% people of color, 15% LGBTQ. So we have a very diverse set of leaders on the platform.

Joe - 00: 07:03: Yeah, it's wonderful. Going after the network effects and getting to the root of like, hey, why has the current management board shaped the way they have? In terms of the network that drives that is that's a really fascinating approach.

Malur - 00: 07:13: And you want to add one more thing is that as we evolve, one of the things we realize that communities are very, very critical in tapping into talent pools. So one of the things we're looking at is how do we expand beyond what we're doing today to other communities? So as an example, last year we hired a CEO for the company. Was a former NFL player. His name is Jared Green, and he's been on board for, you know, about a year now. And we're looking at seeing how we can tap into talent pools as an example. Former pro athletes. There is a statistic that goes, I think 78% of athletes go broke within the first three years. And that's phenomenal. And even though organizations like the NFL Players Association or trust, et cetera, try to help them, even that is a big talent pool of very disciplined in certain areas and have had a great career. How do we get them successful in the corporate life? You know, in sales or business development or starting their own business, et cetera. So that’s our pivot or we're looking at those areas.

Joe - 00: 08:10: Fascinating. Yeah. That is a shocking statistic. I mean, obviously, you know, especially with some of the American pro sports leagues, I had heard of things that happen after they leave the sport and the amount of some of these cases, but I didn't realize the statistics were that extreme. That is really interesting. So, you know, speaking then of talent pools, recently you gave a talk at WilcoHR where you addressed the application of AI in talent discovery, retention and engagement. So you've given an example there of, I guess, the profiles that Strides AI aiming at generally and the kind of approach you're taking to this. But in terms of how the AI-driven recruitment. Tools themselves can help minimize bias and ensure a fair representation of diverse candidates. Can you tell us about how the technology itself is actually helping make these connections happen and making this a fair process?

Malur - 00: 08:51: One of the things we see with AI-based recruitment tools, it can be a double-edged sword, but if you use it the right way, it can help to remove unconscious bias from the hiring process by focusing exclusively on skills, abilities and potential rather than demographics or background. And the key is really in training these systems, using very broad set of data that includes diverse and unbiased data that captures a full spectrum of talent available. So with the right implementation, AI can allow organizations to make data-driven decisions that can give everyone a fair chance. Right? So the trick is really to be able to analyze large data sets, our skills and strengths that traditional hiring practices may miss.

Joe - 00: 09:32: Right. Yeah. That makes a lot of sense. I'm especially the unconscious bias element. You know, obviously there are legal processes and company processes that have people strip as many sources that could tip off unconscious bias as possible and resumes. But I think one that's often quoted is like the name of universities and college degrees. That's something that is part of the job application and yet triggers biases, especially in certain regions. You know, you've got Ivy League, you've got Oxbridge is very heavy associations of big groups of universities. So I can definitely see how that would be useful.

Malur - 00: 10:01: And then yet every day you read all these reports about how there's a huge war on talent or how there's a huge skills gap. It's really the fact that majority of the skills that are required are not readily seen in the talent databases that are available. So how do you bridge that gap? Is the big challenge.

Joe - 00: 10:18: Interesting. So I guess when you're talking about that topic of there's a skills gap, we're ignoring the talent that's there and these unconventional talent pools and things like pro sports players. There's a role there then for AI in. If it's got a data set of, you know, qualified people, it might be pulling out hidden skills and such that, you know, human reviewers aren't necessarily aware of there. Is that the case? Am I extrapolating? I'm playing a bit there.

Malur - 00: 10:40: I think anytime you have large volumes of data, as we do with resumes and people, if a human is doing that, you're going to miss, right? I'm not saying that AI is not going to miss, but at least AI will throw up. It's like fraud detection, right? Fraud detection with your credit card, et cetera. It throws up certain scenarios that could predict that if you are in Barcelona one day and the same day you're in South America somewhere, someone's probably stolen your credit card, right? So same way. I think the AI can detect a certain patterns and saying, these sets of skills or these sets of capabilities lend itself very well to this particular role, which as a human, we can do it, but we're not able to see thousands of resumes in one day, which an AI can do in 10 minutes, right? Which a computer can do. So I think that's really the front end of it. And the second one is there's another statistic I'm going to throw at you. 40% of people that get hired, especially at senior roles, you know, VP level or above leave the company that they joined within for six months. That's because of a number of factors. One, whatever they were told at the big, when they were hired was not what it transpired when it happened or the whole issue of inclusion diversity, right? They were brought in and they were really not onboarded well enough that like they were part of the team. So there's a little number of different factors. So engagement with the employee engagement, the executive equally important once you hire. So I think that whole process, imagine you spend a hundred thousand dollars on a recruiter to hire somebody and then, you know, you got to do that six months later again.

Joe - 00: 12:07: Sure. Yeah. So I think that last point is really interesting. That statistic of people can be sourced and you can have a diverse hiring pipeline, but then the rest of the company of the onboarding process, if you know, their experience doesn't reflect that they're still going to have a bad time. So do you think AI is the sole solution here? Are we still looking at other parts, you know, more human problems within the diversification of these, especially these senior management teams?

Malur - 00: 12:29: Well, the decision in the end has to have a human-centric approach, right? AI is to assist you as a tool to filter it down to a short list, if you will. AI can help in multiple ways, like the platform where we were talking about Strides, which allows you to build relationships even before you have a role. As a hiring manager or a corporate executive, if I build relationships with individuals, I know that six months down the road, let's say I have a requirement for VP of Marketing, I've already met three or four people in that space that I've communicated with, I've talked to, and they know me more importantly. When we decide to hire one of them, it's not a big surprise and there's less likelihood that they're going to leave after six months.

Joe - 00: 13:08: Right. Yeah, that makes sense. That makes sense. You get that support network, you get some pre-experience. It's not, you know, everyone's coming in cold.

Malur - 00: 13:14: Correct. The other thing I want to talk about here is I'm on the advisory board of a company that just recently launched called Latimer, Latimer AI, which is focused on eliminating bias for African-Americans based on Black culture, Black heritage. So there are going to be many, many LLM-type models that come up, which will attempt to reduce bias in multiple different ways. So it's, things are going in the right direction.

Joe - 00: 13:38: Yeah, absolutely. Speaking of, you know, there's many LLMs and one of the points you made earlier was that a large part of how these are going to be successful is going to have a large and diverse training set. One of the concerns that has been voiced about AI-driven recruitment in the past is the training set itself containing some kind of bias. So the training set itself missing out on such a thing. How do you think minority talent can work with the people behind the development of these AIs to make sure that they are found and that ultimately the AI is aware of term, I guess?

Malur - 00: 14:07: Yeah, so the big challenge with LLM is the training, right? As you know, it takes months and months and years of training on data sets. And that's where now there's a technology called RAG, which really stands for Retrieval Augmented Generation (RAG), which is a fix, is something that add-on that you put in that enables you to bring in current data into the mix. So instead of just looking at old trained data using Retrieval Augmented Generation (RAG), you can first look up other forms of data, which is what companies like Latimer, et cetera, use. So that's one way to minimize bias is to have a set of rules and data sets that allow you to reduce the rules based on certain types of bias. The other is obviously making sure that your data sets are based on accurate data. So, for example, Latimer is working with a lot of the HBCUs, a lot of the authors of black literature, to ensure that the data is as representative as possible. And similarly, in other forms of, you know, whether it's healthcare or other LLMs that are being built across the world, a lot of the focus is around ensuring that the LLMs are trained on data sets that are representative of accurate scientific and history and all geographic data.

Joe - 00: 15:22: Yeah, absolutely. That makes sense. To come back to RAGs, just make sure I understand from my slightly more layperson LLM knowledge. So say, for example, you had a model that was trained on a data epoch that ended in 2019. You would be able to update its data, for lack of a better word. You'd be able to use newer data without necessarily retraining the whole model. Is that the case?

Malur - 00: 15:42: That's correct. Because retraining the data is a very time consuming and expensive process, right? It's really an AI framework that allows you to retrieve facts from an existing knowledge base that can augment existing pre-trained large language models to be more accurate, up to date, et cetera. And every day I see new algorithms, new techniques coming up that are helping to reduce bias. I think make sure that. Some of the problems we have with the early versions of ChatGPT, et cetera, are being enrolled.

Joe - 00: 16:09: Yeah. RAG is not one I'd heard of, but I've seen a lot of these new fine tuning techniques, I guess, is what the category of, you know, like things taking these big models and using a smaller data set that's more achievable to everyday companies to specify them. And it's nice to have a specific acronym.

Malur - 00: 16:24: You're right. I think from a corporate enterprise world, the big push is, as you know, many of the companies banned use of the generic version of ChatGPT because, you know, they don't want. Any data to be sent to the public domain. But I think that's where RAG comes in is a lot of the models that are saying, you know, you can now use an LLM base entirely within your corporate firewall that looks at your HR systems or that looks at your benefits or looks at your corporate policies. And now you can type any question and get the answers for that. So those are based on assumptions of LLMs and also using this technique called RAG.

Joe - 00: 17:00: Fascinating. Okay. One of the common threads, especially LLMs, you know, we started talking about their use in the workplace and the places they can offer efficiency and that sort of thing. Is this idea that LLM is going to take our jobs? Now, obviously that's very interesting when we're talking about AI for recruitment, because, you know, it's very much the opposite thing. AI is going to help us get jobs rather than take them. Do you think as a AI company founder yourself that business leaders are approaching this concern in the right way? Like they're doing enough to adequately address the topic of how AI is going to interact for people’s employment.

Malur - 00: 17:35: Probably not. There's a lot of misinformation out there and, you know, Hollywood doesn't help. Every talk that I've given on this topic at conferences, invariably, this is one of the top questions. Is the AI going to take over our jobs, right? But that really is a misconception. And in fact, yesterday at the airport lounge, I met a gentleman who was sitting right next to me. He's from Houston working for a large oil and gas firm. And he asked me the same question. My answer to that was, it's very similar to when computers came on the scene or calculators or mobile phones. Obviously, there's going to be an impact. No one's saying there isn't going to be an impact. But the whole idea is how do you take AI and a combination of other technologies, whether it's AR, VR, robotics, cloud, et cetera, and use it in a way that eliminates the need for jobs that human beings don't necessarily want to do and go up the value chain in terms of the jobs that you want to do and focus more on that. So business leaders, I think most of them are aware that AI is a tool, but it's important for them to get educated about the fact that it is really a tool to augment human capabilities, not necessarily replace them. And the only way you can do that is through education, making sure that every employee is educated on the pros and cons of AI and how to use them and what you can do and what you cannot do. We're still very far away from being sentient as far as AI is concerned. So I think that fear doesn't exist, despite the fact that people claim it could be sentient. It's still a long way. It has to be trained. A lot of it is reflective of the things that we put in the software and the AI itself.

Joe - 00: 19:04: Sure. Absolutely. So you mentioned, obviously, the need for education. I guess this is a question that you are dealing a lot with yourself and your businesses, but also in TCS. How do you go about fighting the stigma against AI in answering this question?

Malur- 00: 19:18: Well, in our case, our organization is fairly tech literate, right? So the majority of our organization are already in the tech business. So AI is not that foreign to us, but we're also dealing with many of our clients who are not necessarily as tech savvy as we may be. So in all of those cases, very clear, transparent communication about the role of AI and how that is going to play in terms of organizational change and how it's going to help enhance specific roles is very, very critical. And then, of course, training and reskilling, whether it's our own employees or employees of our clients, is very critical as well. That has to happen as a way.

Joe - 00: 19:56: Sure, that makes sense. So to jump around a little bit, I'm going back to some things you said when you were talking about RAG. So when we're discussing that, we're talking about the use case of RAG. You said something interesting, which was that it also enables different businesses with different needs to engage in LLM. You mentioned the case of big businesses that don't want to use the public APIs because of data and obviously on the other end of that. There is this idea that small and medium businesses are locked out of this market because you need your own data center with like hundreds of GPUs to train models and such. So I guess to ask the question bluntly, like, is AI just for big businesses? Like, do you need to have OpenAI, Amazon-worthy compute power? Or for our listeners at small and medium-sized companies, how can they get the most out of AI today?

Malur - 00: 20:41: I mean, the short answer to that question is no, they don't. In fact, I would say I think startups and small businesses, they're probably far more nimble and can experiment with it a lot more than big businesses can and rather quickly, right? Big businesses take time to make decisions and they've got to move huge amounts of workforce and they've got a lot of legacy to deal with. Small businesses don't necessarily always have that. And the risk is far less for a small business. And therefore, I think they can look at AI as a way to provide immediate benefit, whether it's in customer support, marketing, customer communication, or even operational tasks, which can be automated. There's so many areas where, I mean, every small business I go to in neighborhood, I live in, they're always short-staffed, right? We all know this. Obviously, that is a huge problem for small businesses because they're not able to find the right workforce with the right skill sets for the right dollar value. So I think AI can play a big role there, whether it's in simple things like bookkeeping or automating certain mundane tasks, routine tasks, improving customer service, enhancing decision-making on non-critical elements, right? That can be eliminated altogether. So leveraging AI can help small businesses become more competitive and efficient. There are so many free tools available today. There's a lot of open-source software available, videos and free courses, online courses available where you can quickly learn about how you can use AI for your business for various industries, whether it's healthcare or tech or retail or whatever it is that you're doing as a small business.

Joe - 00: 22:12: Yeah, I think that point about free tools is really, really important. I think what's especially been interesting since OpenAI, you know, released their API, released the pricing, is becoming aware of how much that's costing all these companies who are basically offering access to the API for free through their own product is, yeah, there's a lot of them. It really is a boom moment.

Malur - 00: 22:30: Yeah, there's a lot of cloud tools. There's a lot of tools available, even just simple things like being able to create presentations, creating graphics, creating marketing graphics, writing contracts. So for example, as a simple example, if you have a situation where you need to write contracts, you know, you can save time and money. Instead of having to go to a lawyer, you can get one of your LLM tools to create a framework for it, and then you modify it yourself, and then you send it to the lawyer, and he or she spends 10 minutes on it and says, “yeah, this is good to go”, as opposed to an hour.

Joe - 00: 23:00: Right. Yeah, we love to reduce our billable hours. So as we get towards the end of our time here, a big part of your work is, you know, on this talent pipeline, on helping unrepresented folks get into these senior management positions. Do you have any favorite success stories of people who have found senior roles through your platform that otherwise would have gone unnoticed?

Malur - 00: 23:18: Yeah, quite a few, actually. I think we've had some success in placing some people on boards that would traditionally have never found each other. Companies would have never found the individual. A specific one was, somebody who had any experience or experience with utilities, and the company was going to be listed in NASDAQ. They were looking for a female board member that had those skill sets. You know, on C-suite roles, we've been able to identify a variety of different roles for people based on our matching algorithm, as well as just getting to know those individuals one-on-one. Then there was another one where we had somebody who had a non-traditional career path, had a fantastic career, both in the public service as well as in private, but was overlooked by the conventional hiring processes. I think a lot of the tools that exist today, you know, the application trackers, the application tracking systems, et cetera, as good as they are, they also create a problem for people with lots of experience who can do multiple things to be able to say, “hey, how do I get shortlisted into one of those?” So I think that's where AI can help as well.

Joe - 00: 24:22: I definitely know some folks who have very broad experience rather than specific, who are a perfectly good fit for a job, but it would require someone to really interpret the broad career that they have to determine that. And they have a lot of issue with those early stages. Those are wonderful stories. Thank you so much for sharing that. And thank you for joining us today. This has been a wonderful chat. And I've definitely learned a lot about this relatively new application of AI.

Malur - 00: 24:46: Thanks a lot, Joe. Really appreciate it.

Outro - 00: 24:50: Beyond the Screen: An IONOS podcast. To find out more about IONOS and how we're the go-to source for cutting-edge solutions in web development, visit ionos.com and then make sure to search for IONOS in Apple Podcasts, Spotify, and Google Podcasts or anywhere else podcasts are found. Don't forget to click subscribe so you don't miss any future episodes. On behalf of the team here at IONOS, thanks for listening.