2808: The AI Imperative: How Safeguard Global is Redefining Work and Leadership

[00:00:01] Welcome back to another episode of The Tech Talks Daily Podcast where every day we explore how technology is reshaping the business landscape and solving real world problems. I'm your host Neil C Hughes and today I've got a great thought

[00:00:17] leader in the field of AI and technology development joining me on the podcast. He's the CTO of a company called Safeguard Global and he brings with him a wealth of experience in developing disruptive AI tools and he's going to be

[00:00:32] sharing his insights on how tech leaders and business leaders alike can make the development process productive, positive and ultimately successful. And if you're just hearing about Safeguard Global for the very first time they're a workforce solutions company adopted and they've adopted a work in any way approach that's

[00:00:52] flexible and people-centric especially in a way that they hire in demand technology talent to achieve business goals. There's so much I want to talk about today about how organizations can better leverage the strengths of humans and AI to create a more harmonious and productive work environment, offer a

[00:01:10] helping hand for any business leaders that want to leverage AI but they're sat on the sidelines wondering what move to make, how to overcome existing technical debt and also how to initiate and sustain a culture chain within an

[00:01:25] organization when adopting things like AI technology and also explore the world of quick innovation sprint. But before we get today's guest on it's time for a quick shout out to the sponsors of tech talks daily and in today's digital age

[00:01:40] where data breaches are all too common securing sensitive information has never been more critical right? Well enter Kiteworks, a pinnacle of managed file transfer security or MFT security. So with its FedRAMP moderate authorization a prestigious certification that they've held since 2017 by the Department of

[00:02:00] Defense, Kiteworks stand as a bastion of security and a sea of uncertainty. This isn't just about meeting standards though it's about exceeding them ensuring your data safety within your organization while streamlining compliance processes that ultimately save you time effort and money. So step into the future of

[00:02:20] secure file transfer with Kiteworks by visiting kiteworks.com today to see why it's held as the most secure MFT on the market so once again kiteworks.com and now let's get today's guest on. So buckle up and hold on tight as I beam your ears

[00:02:37] all the way to Austin Texas where our guest is waiting to join me today. So a massive welcome to the show can you tell everyone listening a little about who you are and what you do? So thank you for having me first off it's an honor to be

[00:02:54] here so I'm Duri Chaitanya I'm CTO at Safeguard Global. A little bit about me I'm a cat herder, problem solver and people amplifier. I support cross functional agile teams at Safeguard Global to ship products that make life

[00:03:11] better for people. A little bit about my journey it's a bit unconventional but I think it's what's helped attract me to Safeguard and some of the things that we're talking about. I've navigated through various different industries and

[00:03:25] geographies through my career so after graduating Boston College I got into I started working at a company that did robotics and PCB manufacturing in New York and then I found myself in product development in China and then I founded a

[00:03:39] social e-commerce company in China as well and then led consulting across several different digital agencies both in China and New York and then I got into the technology transformation really focusing on emerging technologies in ad tech, health care, banking, wealth management in Chicago, New York and the

[00:04:00] UK and for the last three years or thereabouts I've been with Safeguard based at our headquarters here in Austin Texas. For those that don't know Safeguard, Safeguard Global is a global workforce enablement platform. We enable organizations to analyze, recruit, hire, manage and pay anyone around the world

[00:04:22] which means ultimately that they can help make global talent a competitive advantage for their company and our technology platform is really a central pillar of what we are as a company. And one of the reasons that attracts

[00:04:36] me to you guys I want them why I wanted to get you on the podcast today is I always say every day that technology works best when it brings people together and yes this is a tech podcast and what you do is help bringing in tech talent

[00:04:47] but you do it in a way that's flexible and people-centric to get those or to hire those in demand technology talent that businesses need to achieve their goals and it's like that fine balance between technology and people that I

[00:05:01] think is so sadly lacking in a lot of areas. So just to set the scene for our conversation today from everything that you've seen in your work how can organizations better leverage the strength of both humans and AI and

[00:05:15] emerging technology to create a more harmonious and productive work environment and maybe you've got a few specific examples from your experience at Safeguard Global that you might be able to share because we've seen all those bad stories of scary stories that frequent our news feeds but there's a

[00:05:32] completely different side to this tale isn't it? That's right let me start by perhaps setting the scene a little bit yeah because I think it's important to understand directionally where we where we're headed and also where we are with

[00:05:46] AI. So first obviously AI is not new it's been around for many decades I've worked with it in in health care and in banking and an ad tech so you know it's it's it's a technology when we say AI it's a technology that's been around but but

[00:06:06] clearly with chat GPT a little over a year ago something's happened some wave of change has happened and you know just a couple data points that always kind of hit it home for me is that chat GPT was achieved the fastest consumer adoption

[00:06:30] rate of any product ever and that you know another thing that's really kind of sparks an interest for me is that within the developer community it has also grabbed people's attention so I like to go to this site called there's an AI for

[00:06:47] that and within it you can see that there's been more than a thousand new AI projects a month there so that kind of groundswell of interest is is unlike anything we've ever seen before. Now your question is really about how does that

[00:07:08] context kind of work with people you know a lot of people's intuition around AI is that it will replace people because artificial intelligence humans are intelligent therefore they're basically synonymous and replaceable and I think I think the the terms like neural net also kind of causes people

[00:07:35] to believe that these are thinking machines and it's important that people understand that it's not that these these this technology that we have today is not a thinking machine it is a stochastic parrot a very useful stochastic parrot though so what's different about these large language

[00:07:56] model let's take let's take for example you know what we used to do in the past if we wanted to teach an AI to recognize a cat we would we would hell it while a

[00:08:10] cat has ears and a nose and four legs and a tail and it could be fluffy it could be black be white it could be like this it could be like that and you just keep writing all these descriptions of cats until you know effectively you get

[00:08:23] a good description of a cat now obviously that is time-consuming and error prone and hard to maintain and that's what made kind of like the big big data revolution that we were supposed to have over a decade ago kind

[00:08:38] of go nowhere for most companies because it was just really really hard to apply these machine learning techniques of the time now what's different about these transformers is that it's building upon a different sort of architecture it's this deep learning architecture which effectively means you

[00:08:58] can throw lots of information at it and with open AI 3.5 you're able and chat GPT we were able to have both an intuitive interface combined with that large amount of data and and and and statistical model that could then predict the next

[00:09:20] word in this and what you wanted how to say so with minimal amounts of effort you could have human like responses but I hope just I'm being clear though is that effectively what's happening under the covers here is a statistical

[00:09:40] relationship between the prompts that you're providing the context and and the underlying model that's been pre-trained what it's not doing is thinking about what you said and and logically coming up with a response that it's not a it's

[00:09:56] it's not to the point of logic it's just to the point of stochastic parent it's repeating the next thing that should be in line so so how does this work with organizations and people how do we combine the best of both worlds I could

[00:10:15] give one example from from safeguard is a tool we built called chat SG it's a generative AI application rag base so it's it's it's not just going to the large language while that's just doing that but also providing context and then

[00:10:35] it's returning effectively answers about the local HR practices legal stuff that otherwise people would have to go to many different websites different sources to find credible answers for so it's it's hopefully it's breaking down barriers between the knowledge and the response now the question around where

[00:11:01] is the people in that well in the technology there's no people but in the data is a reflection of over a decade of experience in HR embedded into our knowledge base and the constant development of that data by our team of

[00:11:24] over 1800 global HR experts some of the best in the world building that knowledge base and maintaining that knowledge base and observing the answers that the AI is giving back and curating it to make it better and better and

[00:11:40] better but they're developing a generative system by being effectively a part of that system but it's moving them away from what they were doing in the that and that's and that's kind of like kind of a it to me the direction that

[00:11:56] we're headed is making knowledge more accessible more repeatable higher quality more available faster and and it rewards people for their expertise as opposed to just being there to do the work so it's creating the ability to

[00:12:21] have economies of scale for expertise and I think that that's something that we've seen happen in other industries and with this technology it's it's just more broadly possible and I'm so glad you've said that it's so many important

[00:12:38] points there that are often missed when we're talking about AI and also were able to boost a few myths and misconceptions around AI and I think it's so important because many business leaders want to leverage AI now they

[00:12:52] need to be need to know there well they know they need to be doing something and maybe they're listening to this podcast to pick up a few tips on where that journey should start so in what ways do you think things like strategic

[00:13:04] planning leadership development and clear communication play a role in the successful implementation of AI technologies because although we're talking about new exciting technology those fundamentals are more important than ever absolutely I think leadership is a big part of what makes people

[00:13:24] capable of change and capable of seeing the next the next step and as I was saying before I I truly believe we're at a phase transition we've had industrial revolutions of the past which has changed how organizations fit together

[00:13:41] the roles the way in which people operate the IMF today says that almost half of all jobs in the entire world will be affected by this technology and at Davos recently I think I heard it remarked that if Gen AI wasn't part of

[00:14:01] the panel that they would be the outlier though like people are recognizing that this is important no what leadership will help people do is understand how this is different and how they can be prepared and what and the one of reasons

[00:14:22] that is necessary to be explained is because laid large language models are weird they really are a different paradigm something that we're just not used to doing I'll give you an example a lot of people's intuition when some of

[00:14:40] these AI technologies exist is like oh we could just put open AI behind a chat bot and it'll do our customer service for us yeah yeah it's gonna do it's gonna start to respond it's gonna sound human like but it's not gonna give the kind of

[00:14:59] responses that you want why it's because these things like like open AI for instance it has in it the entire intranet and it will respond based on the weights of the information that it received but in the end of the day it

[00:15:17] believes that the world is both flat and round hmm right because on the internet you can find both answers now both answers are not equally correct and probably not in the equal weight so it will on average respond with the Earth

[00:15:37] is round but if you asked it in enough the right way it will eventually give you the statistical response of that is that is in my view wrong and most people to use raw that the world is flat and so this is a kind of an intuition that

[00:15:53] needs to be trained into people to kind of think about of the meta data the the data not just in how the thing is structured not just in the prompt engineering but also in the underlying foundational data and how we can pre

[00:16:15] train and fine-tune these things and and it also gives these leaders an understanding of like why we're not there yet there are certain reasons to be optimistic about this year being yet another sea change that unlocks things

[00:16:33] for many organizations but they must first understand kind of the key challenges that prevent to prevent these this these technologies from being taken at completely freight face value it like it at first blush your intuitions

[00:16:47] will be wrong and another issue we've got to talk about of course is technical debt before we start adding new shiny technologies into the mix and you have indicated that legacy systems often create data accessibility issues for a

[00:17:04] li implementation so is any advice you can share on strategies for reshaping that technological architecture to facilitate that better AI integration and stop problems before they occur so I could offer what I've seen work yeah for

[00:17:20] organizations these days at at safeguard we've applied an approach we call well that's called the industry data mesh this became kind of popular about two years ago and the concept it is is fundamentally about ensuring liberation

[00:17:42] of data from your applications that you can lay let it to be independent so that it can be then leveraged for these kinds of purposes and it's about a democratized data governance data literacy and quality programs that that

[00:18:10] can be operated at scale without a central team now this is all kind of nice to say if you're on greenfield but as you say there's this kind of a lot of legacy out there and organizations that have some of the most valuable data in

[00:18:32] the world might find it difficult for them to leverage it to be able to win in this new AI race some of the things that worked for safeguard is that we made it a common like a top-level priority to provide to provide teams the

[00:18:53] training and the governance structures and the incentive structures to make data a first-class citizen that we treat it with a lot of safety protocols to make sure that we respect our clients privacy and security and then we think

[00:19:13] about optionality though we don't design data to be leveraged just for one use case but instead design it to be available within the system which is a big part of the concept of data mesh and even with the best plan in place they

[00:19:30] also need the right people so there is a huge tech talent shortage at the moment a lot of AI expertise is in demand so how do you address that talent gap when it comes to AI expertise especially considering that required skill set that

[00:19:45] sits between data science and software engineering any advice that you'd offer to tech leaders or business leaders that are facing challenges around securing the right talent right now I think the skill that's most in demand right now

[00:19:57] that's needed is this kind of AI engineer now what that AI engineer is it's not building foundation models they don't need to necessarily even be you know have a have a PhD in artificial intelligence instead what they need to

[00:20:13] be is a problem solver that has experience across a broad array of the emerging field and the ability to do something called technology scouting to find the best-in-class open source capabilities and apply it in a way that

[00:20:29] is evolutionary meaning that you know they're not just thinking of it as like this is the solution but instead think of it as a door to walk through that then opens up other doors so how do we go about finding that talent well it is a

[00:20:47] big challenge as you said there is a talent shortage and often the best person for the job isn't working within 50 miles of your office and so I'd say safeguard I've very strongly pushed the premise that you find the best talent

[00:21:05] anywhere period and then you work in any way to make it make that person a part of your organization so for instance at safeguard we have 155 product and engineers across 22 different countries and many times we have teams of eight

[00:21:24] people that can be spread out across three or four continents that's um that's quite a different sort of structure that I would have imagined 10 years ago but it works why because you're emphasizing skills and talent

[00:21:42] over an experience over proximity to your office space and in and when you put it like that when I put it like that it's kind of obvious but still people tend to still lean into old old patterns because they think that that gives them

[00:21:59] more control in my room and my experience this sort of field this kind of challenge that we're dealing with the complexity and interconnectedness interdisciplinary aspects of the economy that we're working in means that leaders need to relinquish control and instead double down on a strategy that focuses on

[00:22:21] the individuals on the ground being the smart ones right giving them the autonomy and the capabilities to go and run out of themselves and so you have to basically make and pull out all the stops for your talent strategy beyond that one thing

[00:22:39] that we talked a little bit about before was the importance of leadership even with a lot of experience and skills these large language models and paradigms are still emerging and no one is an expert in it yet and so leaders

[00:23:02] need to create space for people to experiment they need to be able to trial things they need to they need to pull in technologies that are in open-source projects that have only been around for three months and are maintained by 15

[00:23:18] people like that like that would have been unheard of when I was working in banking but I see it as a key to our strategic advantage is that we're willing and able to try things and and let it say and if it fails stop doing it

[00:23:37] Chattisji is an example we rebuilt it three times before we ever released it and by the way we released it in three months with less than three people so three people in less than three months rebuilt the same application three times

[00:23:53] so I think it's a it just goes to show you that experimentation and trusting the people on the ground is really the remedy for the complexity that we have and if I look back at my own career in IT a lot of the same problems are always

[00:24:08] seem to repeat over the years and that is you can have the right people in place to deliver that technological change you can have the latest must-have technology that's going to solve all the problems but it's no good if you're gonna got

[00:24:22] that attitude of but hey we've always done it this way so how do you initiate and sustain a culture change within an organization that is maybe resistant to adopting new technologies like AI are there any management strategies that

[00:24:36] you found particularly effective in guiding people through that change curve it's the oldest tale of time isn't it but possibly one of the most important ones so I'll put my hand up right now and just say this is probably the

[00:24:49] hardest part of the job yeah the culture gap because an expert put particularly if you've hired a lot of really smart experts will reflect on their past and say hey this is work for me this is what's gotten me here and now you're

[00:25:08] saying we need to go try something different and you know so so you are asking people to do something that is instinctively hard to do and it comes with you know ambiguity and you know concern that we don't know enough to

[00:25:24] actually make this decision so a couple a couple suggestions I would make is there's a great author on the subject Cotter and he's come up with a cotter change model that one of its key components is the idea of working across

[00:25:45] the organization not just within your silo of your department changing minds by helping solve a real foundational problem that they have within their within it would that they have within their daily work and convincing people that you're

[00:26:04] not going to just automate away their jobs because in the end of the day even software engineers today are worried that well with this new technology am I even relevant anymore you have to really lead with value that says I'm not here

[00:26:21] to replace people with technology I'm here to build the best organization possible and that means that we're gonna make you the best at working with technology and so I think really leaning into those guys and believing don't make

[00:26:39] them just a marketing banner but make them really core to how you approach problems it's more likely than anything to get your to get people people to buy it I suppose and I had Dave West CEO of Scrum on last week so another question

[00:26:58] that I've got to ask is can you elaborate on the concept of quick innovation sprints and how those sprints might have led to effective proofs of concepts that and maybe some lessons have been learned through this agile approach because again incredibly exciting approach to innovation isn't

[00:27:17] yes Dave is brilliant I he gave me some help a few years ago when I was actually exploring some of the similar concepts and I think for us one of the keys is to have low-cost low-risk investments to bring down the fear and to educate

[00:27:43] people so you know this is what we call the innovation sprint you know it's very commonly known as design sprints it's a say it's the idea that you try to solve for complexity by learning through experience the rat any breakdown

[00:27:59] people's preconceptions before they've calcified that's that's a technique that works very well particularly if you make it not just an engineering or product experience but you also involve your stakeholders so people like operations and my in my situation people like operations marketing and sales are

[00:28:25] often we see them as a key component of the overall system and so involving them so that they're also feel part of the experience of learning and creating the solution that that's what makes it successful and then on the back of that

[00:28:44] innovation sprint now you're ready to kind of play right now you're ready to kind of establish with your initial hypothesis is your MVP and still treat it as a you know short-range you know high risk but potentially high reward

[00:29:03] activity but with perhaps a bit more weight behind it and you slowly and you slowly work towards a state where you've met products market fit and it's not forced here it's not a forced assumption you've really gone through that diverge

[00:29:21] and converge activity in the innovation spread India and the succeeding MVPs and I also think bringing AI into an organization it is more of a journey than a destination it's not just a box ticking exercise so what does a

[00:29:37] commitment to continuous improvement look like in the context of AI development are there any key KPIs and or the organization should be monitoring key KPIs for how the organization is performing with AI are I would say things like your traditional business ROI measures like ultimately bringing us

[00:30:05] back to creating true value because though the challenge that I remember of the big data era was that we spent I don't know how many millions and every organization invested in like getting their data into one place but

[00:30:20] then we couldn't figure out how to get like true disruptive value out the other end and so for me I really focus on small investments that have laser focus on developing a a valuable outcome and being really specific about what that

[00:30:42] outcome is so we're not navel gazing on technology but we're really focusing on you know let's let's improve this customer satisfaction metric or let's let's increase the organization's ability to make good decisions by surfacing information earlier those kinds of things you can you can start to

[00:31:02] put contextual KPIs around and I think that it gives you the best success when when you can show the organization something on the balance sheet that you've achieved. Brilliant absolutely love that and as I said a few

[00:31:18] moments ago in our interview here many business leaders will want to leverage AI maybe they're listening to this podcast maybe looking to pick up a few tips of where their journey should start and learn from someone like your good

[00:31:28] self here that has walked a mile in these shoes already so any advice on how people listening can incorporate creative problem-solving techniques into the development of new AI solutions and any examples of you can share of where creative problem-solving might have led to breakthroughs or success in AI

[00:31:47] deployment I appreciate it's probably not too much you can share but is anything you can? Well first of all I think small teams of the best talent is the first and most crucial ingredient that if you as an executive set up a

[00:32:04] team that doesn't have the skills experience or you know feeling of empowerment to go and make wild decisions then you haven't put them in a place where they can be creative that ultimately they're just going to go and

[00:32:24] paint by the numbers so the talent the people putting them in the right place is number one number two is make sure that they're really invested with the right idea of the KPIs are going back to the last question around what is the

[00:32:38] outcome that you're looking for them to achieve third is I would say make sure that they have freedom to operate so at Safeguard we really believe in work in any way which means that they are responsible for choosing their tools

[00:32:52] they get to pick the technology they get to go and do the technology scouting and you know you're constantly coaching them to say but if it's gonna take you a month what would it take to do it in a day and so you're pressuring them to with

[00:33:11] constraints around time to think differently right so that constant compression of time is also a really great asset you will hear anybody that knows me walk and say like they might say something nice but usually I'll end

[00:33:29] something with but he really really wants it now and I think that I think that in some cases that could be a very bad method pressuring people on time could cause burnout and all kinds of bad things but in the right circumstances

[00:33:45] with the right values behind it and the right flexibility and the ability for people to push back and say it's not possible it can cause people to think differently it could cause people to squeeze out waste from the system and to

[00:33:59] focus in on their key assumptions and that's often where insight happens oh and one more thing I can't overstate that ultimately an engineer is only as smart as his or her experience if you can combine that with other disciplines other

[00:34:20] people in the organization that have different experiences then you're much more likely to find some unique and uncommon insight and you see this often in in in the ecology is that the place in the world where there is often the

[00:34:40] most biodiversity is often on the borders between deserts and rainforests it's like the it's the places where two different systems merge where you get this groundswell of beauty and and an uncommonness and and bizarreness and and

[00:35:00] that is what I believe leads to innovation that's a beautiful moment to end on but before I do we started the podcast talking about your origin story and what put you on this path but as we come full circle now if you look back on

[00:35:17] your career what was the funniest or most interesting story that has happened to you because I suspect that you picked a few up along the way right I certainly have so at the beginning of the pandemic I I I'd I'd been working at a wealth

[00:35:38] management company in in New York and I had I'd been I had to be in formal clothes when I went to the office and I said to myself when the pandemic happens that I wonder

[00:35:57] if I'm ever going to wear those clothes again it's now been how many years and they're still gathering dust in my in my closet I think I think it's it's just a wonderful thing

[00:36:08] to be at home and express myself in the way that I want in my hoodie and my t-shirt and and still feel like I could do a great job and I think it goes to I tells it show it goes to show me that

[00:36:24] you know I found a home here at safeguard because you know everybody's at the same kind of way we're a remote first company and I haven't had to wear a suit since yeah absolutely love it I'm exactly

[00:36:36] the same and if for anyone listening just wants to find out more information about safeguard global dig a little bit deeper on some of the topics we explored today where would you like to point everyone listening and how can they contact your team so go to safeguard global

[00:36:51] comm you can contact contact us and you could also try a lot of our tools for free there also if you want to get in touch with me directly you can find me linkedin or on X at derecha tie well I

[00:37:07] will add all those links to the show note so people can find you nice and easily and although we've talked all about artificial intelligence today and how to implement it into an organization for

[00:37:17] me the big key to our conversation is about people and how the emergence of AI is not about replacing people it's about providing an opportunity to reassess work dynamics and prioritize human strengths like creativity strategic thinking emotional intelligence all those areas where

[00:37:35] automation falls short but just a big thank you for your help today and bringing this topic to life thank you now great to be here now that concludes our conversation today and I think as

[00:37:46] we explored the world of AI development and the opportunities that it presents for tech leaders I think the big takeaway was things like quick innovation sprints continuous improvement and creative problem-solving these are the key factors here and the emergence of AI is indeed reshaping

[00:38:03] the way we work but it's vital for companies to invest in their teams and their people and only then can you adapt to these new technologies but obviously I make it sound so easy don't I but of

[00:38:15] course there are so many challenges to overcome from data accessibility to talent acquisition to fostering a culture of change and these problems and these challenges are something felt inside of every organization of every size in every country that I know you're all listening

[00:38:32] in today so you've heard from me you've heard from today's guest it's love to hear from you what are your thoughts on integrating AI into your organization have you faced any challenges

[00:38:43] that we've discussed today and how did you overcome them whether you want to leave me a quick voicemail by hitting record on your whatsapp and emailing me to tech blog writer outlook.com or maybe

[00:38:56] instagram twitter linkedin you can send me a message there equally if you'd like to have a conversation with me come on the podcast and share a different side to this story I'm up for

[00:39:07] anything so please keep your messages coming in thank you for tuning in today remember the future is dominated by tech but the future is also collaborative so until next time keep innovating keep exploring the endless possibilities that technology brings and we'll reconvene bright

[00:39:26] and early tomorrow morning see you then