Episode 11 - AI Deployment at Multinational Organizations

00:00:18:01 - 00:00:45:18
Jie Ren
Hello everyone! Welcome to my podcast, When Tech Meets Ed. This is Professor Jie Ren again. Today, I have the great pleasure to be speaking with my friend Jorge Martinez Navarrete from the United Nations. And he's the information technology officer there. Today, we'll be speaking about a very important topic, which is AI deployment in large organizations. And he's speaking in his personal capacity.

00:00:45:20 - 00:00:51:14
Jie Ren
So Jorge, thank you so much for connecting with us from the UN. Welcome.

00:00:51:16 - 00:01:04:03
Jorge Martinez Navarrete
Thank you, Jie, for having me here in your podcast today. I'm glad to be here and, yes, talk about my experiences. And as you said, I'll speak on my personal capacity.

00:01:04:05 - 00:01:20:09
Jie Ren
Absolutely. Yes. Jorge will be speaking in his personal capacity. Alrighty, so the very first question for you is really about your personal journey pursuing tech. Do you want to introduce yourself, and also maybe with an emphasis on that?

00:01:20:11 - 00:01:49:20
Jorge Martinez Navarrete
Yes. So, I have been a UN personnel for over 20 years now. And my experience has been in the ICT information and communications technology field. I've been deployed to different duty stations like the Democratic Republic of Congo. I've also been deployed to Bangkok, Thailand, Lebanon, Jordan. And my base is New York.

00:01:49:22 - 00:02:30:16
Jorge Martinez Navarrete
And throughout those experiences, I've been focusing on implementing information systems, designing and implementing information systems for all of the different domains in the United Nations. So we, at the United Nations, we do a whole bunch of different things, from economic research, technology, emerging technologies, and how those would apply to society. But we also do very mundane things like ensuring that we have safe telecommunications for our country offices, that our personnel have information systems to manage our financial assets and our personnel and our human resources.

00:02:30:18 - 00:02:56:17
Jorge Martinez Navarrete
So, and we also coordinate fleets of vehicles for our peacekeepers in peacekeeping missions. So that includes tanks and aircraft, helicopters. So it's really the whole spectrum. So yeah, I'm very excited to talk about how we are now, dealing with the implementation of AI, which is, it's a big shift for many

00:02:56:19 - 00:02:58:05
Jorge Martinez Navarrete
requirements.

00:02:58:06 - 00:03:19:05
Jie Ren
Yeah. I'm so excited about the whole spectrum of different types of technologies that you just mentioned. That could be totally different than the cockpit world. But before we go there, I do have a question which is, about your personal journey. Like, what first drew you to tech? Maybe a little bit of childhood story here and there?

00:03:19:07 - 00:03:52:14
Jorge Martinez Navarrete
Yeah, I mean, I'm a mechanical engineer and an industrial engineer, so I was always interested in productivity, efficiency, do more with less resources and automating things. So, from very early, you know, when I was in high school, I was discovering the internet and everything we could do with internet. So I was never even though I'm a mechanical engineer, I was never– I never had a job designing engines or parts.

00:03:52:16 - 00:04:27:18
Jorge Martinez Navarrete
I was always working in how IT, how information systems could be used for manufacturing or for production. That's when I switched to industrial engineering and, my first, professional jobs were in the car industry doing e-commerce for vehicles. So how we could implement systems to liaise our factories with our suppliers to reduce the time from, in order to, getting parts and then assembling faster using automation and information systems.

00:04:27:18 - 00:04:54:16
Jorge Martinez Navarrete
So all my life has been automation. After the car industry, I moved on to the retail and e-commerce, 100% e-commerce. And this is when I moved to I started my career in France. And then,I moved to the US. And in the US, I was working for e-commerce retailers on their e-commerce deployments.

00:04:54:16 - 00:05:17:06
Jorge Martinez Navarrete
This was where e-commerce was taking off. And, because– ,early 2000s, and because I was doing e-commerce and logistics for retailers, that's when I joined the UN because the UN has a big– we run a real big and complex logistics system with all of our country offices and also our field operation.

00:05:17:06 - 00:05:46:05
Jorge Martinez Navarrete
So we have, large amount of peacekeepers deployed in places where there are no distribution systems that you can just call up. So we have to manage our own distribution chain for food, for fuel, and for assets like cars, generators, everything that you require to maintain a peacekeeping operation in a desert or in a post-disaster location.

00:05:46:07 - 00:06:04:19
Jorge Martinez Navarrete
You need to bring everything with you, from water to antennas to radios to tents. And, so that's how I joined the UN. And so since then, it's just been a never ending, trying out new technologies to solve our problems. So, yeah, it's been a wonderful, wonderful experience.

00:06:04:21 - 00:06:28:15
Jie Ren
Yeah. I really like the word that you use about never ending, because we see so many waves of technological advancement. And then the organizations non-for-profit, such as the UN and also a for-profit, such as the companies that you are working for in the car industry. Right. Trying to embrace the different types of technologies, right, for higher efficiency, higher productivity, etc..

00:06:28:21 - 00:06:57:11
Jie Ren
So let's go back to tech deployment first, generally speaking, to set the stage, right. So, over the many years you probably have deployed many types of tech systems, right, in different organizations. So do you want to give some examples? I know that you talked about logistic systems, about the distributions, right, of the essentials. Then any concrete other examples you want to share?

00:06:57:12 - 00:07:03:23
Jie Ren
And then also what do they have in common and any specific unique challenges.

00:07:04:01 - 00:07:35:02
Jorge Martinez Navarrete
Yeah. So yeah, thanks for that question. So I think I want to highlight two different scenarios that are very, very different. So the first one is an internal logistic system for managing our own fleet of aircraft. So in some peacekeeping missions, we have up to, 200 or 300, combine aircraft between helicopters and airplanes, which we use to move personnel, but also to move assets.

00:07:35:04 - 00:08:00:00
Jorge Martinez Navarrete
So I work with our logistics department, the military contingents to design a system to track our own fleet and issue tickets to passengers and labels for moving cargo. So this is a very, it's something that anybody in the private sector relates to because it's very similar. Yeah. Logistics. Right. So that's a very, mainstream, type of system.

00:08:00:02 - 00:08:29:01
Jorge Martinez Navarrete
And on the other hand, at a point when I was in Thailand, I worked with ministries of information technology in four countries at that time was India, Bangladesh, Nepal and Bhutan. And they were focusing on how to leverage the internet to train their teachers so they could then in turn train their people, their children, better nationally.

00:08:29:01 - 00:09:06:03
Jorge Martinez Navarrete
And they wanted to deploy e-learning systems to reach out, you know, hundreds of thousands of teachers in a country like India. And so they were working at the time with the UN to see how other countries were using e-learning to quickly train a lot of trainers, very fast. And so that was amazing because we were piloting the e-learning platforms and at the same time, putting network connectivity to rural villages that didn't have connectivity.

00:09:06:05 - 00:09:20:05
Jorge Martinez Navarrete
And so we were working with the governments of those four different countries. So it was very challenging. It was very exciting going to these new telecenters where we would install a new antenna and then train people on how to use the system.

00:09:20:06 - 00:09:36:07
Jie Ren
And it's like e-commerce, but in the context of learning. Right. And also it's not only about one organization per se, it also involves the local infrastructure. I just said, right. So the challenges it could be more massive than the corporate world. Yes. Please continue.

00:09:36:09 - 00:09:44:02
Jorge Martinez Navarrete
No, no. Yeah, exactly. So that was a challenge interacting with the different parties, the local stakeholders.

00:09:44:04 - 00:09:44:21
Jie Ren
Yeah.

00:09:44:23 - 00:09:47:01
Jorge Martinez Navarrete
It was a really exciting project.

00:09:47:03 - 00:10:15:11
Jie Ren
Nice. So, maybe let's shift a little bit to the internal policies, right? That, very much about tech governance, like, internally. Is there any common theme related to that? And then over the many years, do you see any changes in terms of the internal policies, especially when you are facing different types of challenges, in terms of different types of technologies and in different types of contexts?

00:10:15:12 - 00:10:17:08
Jorge Martinez Navarrete
Wow. Yeah, that's a really good question.

00:10:17:10 - 00:10:18:21
Jie Ren
That’s a really good question.

00:10:18:23 - 00:10:56:12
Jorge Martinez Navarrete
I mean, it really that's– I think there's a it's– you're going to a very interesting topic because, from my experience, ICT, like governance and corporate governance of technology has been a continuously evolving field. So we've gone from where people were given laptops for personal productivity and you would have 1 or 2 corporate systems, to where you have absolutely 100% deployment of IT in every department.

00:10:56:14 - 00:11:32:16
Jorge Martinez Navarrete
And then allowing people by themselves to build their own databases, their own, intranets for sharing information, their own chat groups. So that completely changed how we had to control that growing ecology inside of the organization. And so once that was sort of under control, then you had the mobile revolution. So then people started leaving their office doing everything from mobile phones that were not governed by the organization, but they were the property of the employee.

00:11:32:18 - 00:12:00:03
Jorge Martinez Navarrete
And everybody wanted to access the corporate systems from outside of the office. So that generated a new transition. And then, currently, I believe that with the growth of AI, now we're in a new transition where now we're trying to see how do we govern AI, and how do we make sure that all of our staff are using AI because we want them to use AI in a responsible way?

00:12:00:05 - 00:12:12:06
Jorge Martinez Navarrete
So there's a new trend to it, to how to adapt our governance models, internally. And I know for a fact that every organization is going through the same struggle.

00:12:12:08 - 00:12:43:05
Jie Ren
Yeah. I really like the word that you used about governance, right. So tech governance right now is AI governance. Of course, there could be some confusion. And also it's very customized to every single organization for sure. We'll be talking about that very soon. So now let's move to AI. Right. So in terms of– I am sure that you have been deploying AI systems in your organization.

00:12:43:07 - 00:12:54:08
Jie Ren
And then first of all, let's talk about the benefits of doing so. So from your observation, what are the benefits? How do they change the day to day work, the AI systems?

00:12:54:10 - 00:13:20:23
Jorge Martinez Navarrete
Sure. I mean, first I want to mention a little bit about my role in the UN. So I am, I lead the innovation unit in the technology department, and I'm also in the AI enablement team, which is an ad hoc team that we formed to quickly bring in AI to our day to day work for all our personnel.

00:13:21:01 - 00:13:51:01
Jorge Martinez Navarrete
And so, yeah, so we've been tackling these more or less in two, in two pockets. So one of them is empowering personnel with individual tools so they can just become more productive, at the personal level. So this is equivalent to, you know, rolling out AI agents, AI assistance that the staff can configure by themselves.

00:13:51:01 - 00:14:29:16
Jorge Martinez Navarrete
And I'm not going to name any brands because we're not here to promote any specific company. But yeah, so giving personal AI productivity tools for staff to use them in the best way they can, that's one angle. And then the other angle is, incorporating AI or enhancing our corporate systems with AI tools. Imagine business systems where we are processing approval of for example, forms, payments and individuals looking at forms and checking whether something matches.

00:14:29:18 - 00:14:54:22
Jorge Martinez Navarrete
This has been automated for a while with, you know, databases and programming, but now we can do more like recognizing images or making some simple decisions with information systems. So that second part is more complicated than the first one. In the first one, you just buy tools and train people and how to use them and make sure that they're not introducing risks.

00:14:55:00 - 00:15:28:23
Jorge Martinez Navarrete
But the second part requires more of, seeing how your systems are going to be transformed, which, when it's worthwhile to put the effort into, upgrading a system versus maybe retiring some systems and adopting entirely new systems. And also, this is where the big opportunity lies, because this is when where you can really do AI– I mean, it used to be called digital transformation, and now it's a transformation powered by AI.

00:15:29:00 - 00:16:06:12
Jorge Martinez Navarrete
So this is where you can make a big, big change. And so it's also the most challenging, in terms of what value, what value these AI can bring. So definitely at the personal productivity scale, there are a lot of industry studies that show how personnel are increasing the productivity. And just anecdotally, we can many of us that use AI daily, we notice how we produce some content faster, clean up our content in better ways.

00:16:06:12 - 00:16:40:23
Jorge Martinez Navarrete
So that's the version of productivity gains. At the enterprise level. It's a little bit hard to measure. But, you can when you modify certain systems and you reduce some steps, automating them with assistance of AI you can measure if it's a process that is done a thousand times per day, and you have a 100,000 employees taking that little, saving one minute there, you notice the I mean, the time savings, aggregated time saves are incredible.

00:16:41:01 - 00:17:01:02
Jorge Martinez Navarrete
But, that doesn't mean that, if I save a minute in my life completing a certain form online, that doesn't really make me more productive. You know, it just saves that one minute, and it just sort of improving my job satisfaction because I don't have to spend that extra minute filling out that form.

00:17:01:04 - 00:17:15:07
Jorge Martinez Navarrete
But the challenge is how organizations can convert this time into, you know, better products, better research, better service to your clients. So it's a discovery mode.

00:17:15:09 - 00:17:37:10
Jie Ren
Yeah. I feel like, actually, there are two aspects of topics that you have mentioned in your answer, if I may categorize them. So the first one is the categorization of the two different levels of impact. One is on the individual level, right, is about personal work. Let's not talk about productivity per se, because that's about the second aspect that you mentioned.

00:17:37:12 - 00:18:03:16
Jie Ren
And then the other one is on the enterprise level, the infrastructure level. And then you use the really wonderful phrase to describe it, which is digital transformation. Of course, we know digital transformation for so many, so many, so many years. And then every time there is, you know, like a new wave of technology and then that means there would be a new set of opportunities to transform businesses, maybe to embrace new business models, etc.

00:18:03:16 - 00:18:34:17
Jie Ren
And that is the most challenging part, especially the transformation is affecting every single person in the organization, which I will go to in terms of adoption of technology very soon. And then the other part that I really, really want to mention is the productivity on the individual level, as you said, that saving that one minute, right, may not mean that you increase your work productivity again, that is really related to the workload that is being allocated to you.

00:18:34:19 - 00:18:53:17
Jie Ren
Like maybe it's a personal choice in terms of how much work you want to do. In many cases, it's not a personal choice. It is kind of like you get this workload from your boss, right? And then again, that needs a– from the organizational level in terms of how to distribute this work right to the individuals as well.

00:18:53:17 - 00:19:24:14
Jie Ren
So that itself is a big, big topic for sure. Right. So then let's shift our conversation a little bit to the adoption of the technologies, especially the enterprise level of impact when you are trying to deploy the AI systems per se. For example, I'm sure some departments could be very embracing the technology, and some departments are like reluctant.

00:19:24:16 - 00:19:28:17
Jie Ren
Right. How do you balance such tension?

00:19:28:19 - 00:20:00:08
Jorge Martinez Navarrete
Yeah, I think a– that's so, you describe it in the way that I experience it. We have departments that really are interested and eager, and then others that are very reluctant and even within departments, you have also both types of people. From my experience, what we found is that, people, the more informed they are, the more– the better the discussions can be.

00:20:00:10 - 00:20:24:00
Jorge Martinez Navarrete
So sometimes there is initial reluctance and once you qualify the scope of the implementation and people say, yeah, that's okay. You know, we are afraid that– we read, for example, that AI can hallucinate. So therefore in our type, in our field of work, when you speak, for example, to lawyers, they would say, you know, in my field of work I cannot have a single mistake.

00:20:24:02 - 00:20:44:19
Jorge Martinez Navarrete
And that makes perfect sense. But when they know that, well, we will use AI, to prepare initial drafts and that this is going to save you a lot of hours of your day, you know, copy pasting content from different contracts. Then they say, this is wonderful. So as long as the scope of the deployment is contained to that, they are very happy.

00:20:44:19 - 00:21:15:18
Jorge Martinez Navarrete
So, it is a process where all the potential owners of solutions need to really dive, deep dive into what the technology can do, and then they can pick and choose what's appropriate for their needs. And as long as we leave them the choice to set the bar to how much they want to incorporate, then they are happy with it, they welcome you.

00:21:16:00 - 00:21:25:22
Jorge Martinez Navarrete
But if you come in and push in with an attitude that you know their business better than that, they do, and they will immediately push back and say, no, you know, get out of my office, please.

00:21:26:00 - 00:21:56:05
Jie Ren
Yeah. So very well said. Right. So it truly is about the alignment, right. You want to demonstrate to them that these systems can truly help you in terms of your workflow, etc., and also maybe, right, encourage the early adopters of some of the technology and let them set the examples, right. And then if they can gain the benefits, from using AI for example, and then, you know, the other employees, they are like, oh, maybe, you know, I can use AI systems as well, right?

00:21:56:07 - 00:22:20:15
Jorge Martinez Navarrete
Yes. And I think that there is also in every decision making, there are some like internal influencers they may not be the director. They may be somebody who has no title. But when they're deciding whether you adopt a technology or not, everybody looks in a certain direction and is that person who, if they would say, yeah, this is good, then it happens.

00:22:20:18 - 00:22:27:16
Jorge Martinez Navarrete
And if they're like, no, no, this is not good. So you have to be mindful of identifying who are the technology influencers.

00:22:27:18 - 00:22:56:12
Jie Ren
That's great. So, maybe let's shift our conversation a little bit to tech governance. And here AI governance. Right. We know that many companies or organizations are very much trying to deploy AI but there is a very big concern which is data privacy. Right. Because part of that truly is about sharing your internal data or maybe part of internal data or a lot of internal data with third party vendors.

00:22:56:14 - 00:23:00:17
Jie Ren
How do you address this data privacy issue?

00:23:00:19 - 00:23:26:08
Jorge Martinez Navarrete
Yeah, I mean, so I think that, I see people who have been doing technology governance are very comfortable with discussing these because just a couple of years ago there that the magical world was doing data governance. So, they were going through the struggles of, well, should we put our data in the cloud and what are the controls that we need to put before we put our data in the cloud?

00:23:26:08 - 00:23:50:19
Jorge Martinez Navarrete
And should we separate, business critical systems that can go to the cloud and those who have to be on my premise, super close, cloud that I manage and that it's logged with triple triple locks. Right. Only the CEO can go into that data center. So I guess for the– in the technology department it's not a new thing.

00:23:50:21 - 00:24:19:01
Jorge Martinez Navarrete
But I think that for the business, for the business users, they read in the newspaper that. Well, yeah, my data is going to a third party AI company, and I know that they are training their models on my questions and my answers, my data. They said I don't want my data to be used for training, because once it's used for training, maybe available to another company.

00:24:19:03 - 00:24:50:06
Jorge Martinez Navarrete
And this is right, like this makes sense. And it's very important that the business users understand this. And I think as this discovery of AI and how it works, sinks into the, into the brains of all of our, of our decision makers. They have also appreciated more the importance of having control of your AI for certain practices.

00:24:50:08 - 00:25:16:22
Jorge Martinez Navarrete
So, that the necessary effort and expenditures that you need to put into having your own AI capacity in-house, being able to run it in your own servers with your databases where all the data is controlled internally and for your business critical operations, making a business case that it is justifiable to keep that AI in-house.

00:25:17:00 - 00:25:33:09
Jorge Martinez Navarrete
Not for everything, but for the things that are very critical to your business. And so it is a challenge because it is much easier to just call a third party vendor and say, give me a solution. And they will say, yes, of course, the solution is ready to go tomorrow and is hosted in my cloud. Right?

00:25:33:09 - 00:25:53:12
Jorge Martinez Navarrete
And so you're like, okay, well, yeah, okay, I will put it there. But I need to– I need guarantees that all my data is safe. Right. So that's and that's where you start saying, well how much can they guarantee. And depending on your industry you have different levels of tolerance. So it's okay. It's much more efficient to keep data in the cloud with third party vendors.

00:25:53:12 - 00:26:12:14
Jorge Martinez Navarrete
So it's perfect for, you know, 95% or 99% of the use cases. But for some use cases, you would be inclined to keep your AI in-house. And that brings a whole other bucket of technologies that you need to learn and things that you need to think about.

00:26:12:16 - 00:26:16:16
Jie Ren
Yeah, like the Open Source AI. Right. Do you want to talk about that?

00:26:16:18 - 00:26:40:13
Jorge Martinez Navarrete
Yes. Yes, exactly. I mean, I think that something that a lot of the people working in the, in the space are very aware of are two trends, related to Open Source AI, on how Open Source AI is growing so fast and competing very closely with closed proprietary AI models, for example, language models, for example, but also, small language models.

00:26:40:19 - 00:27:20:22
Jorge Martinez Navarrete
Yeah. How having the ability to run AI inference, I mean, the products of what AI models generate in local devices, in telephones, in personal laptops, without having to go to a third party. Yeah. And for at least these two trends are very– moving very fast. And I think companies are like, private companies are leveraging this for running their own AIs in-house and also to build, build AI into, standalone products like, robots that move around.

00:27:20:22 - 00:27:30:14
Jorge Martinez Navarrete
And they don't have to be connected to the internet. They have their own local, and AI capabilities. And, yeah.

00:27:30:16 - 00:27:44:17
Jie Ren
Yeah. So from your own observations and also your own usage, which one is better: open source and then close one. I know this is a big question because either one could be better in different contexts.

00:27:44:18 - 00:28:15:02
Jorge Martinez Navarrete
I mean, I think they're both necessary options to have in the market. In terms of benefits to society and to competitivity, to having open source AI is super important because you can have researchers all over the world share among themselves how to build AI because– selective intelligence.

00:28:15:04 - 00:28:44:02
Jorge Martinez Navarrete
And also it's an equalizer technology. You can have a small startup in a developing country, you know, use very advanced tech, and then they can build a really good product and then compete with companies that are in the very developed countries. So you're equalizing the playing field. And also it makes sense for companies to make products, make a profit of packaging products that they can sell and high quality, high-end products.

00:28:44:05 - 00:28:58:02
Jorge Martinez Navarrete
So both, both should exist. But it's great to have the option and access to open source technology to learn and to research and to build new products as well.

00:28:58:04 - 00:29:27:18
Jie Ren
Thank you. So I really enjoyed our conversation. I do have one final question for you. Since again, this podcast has Ed in it. So usually I will ask one question related to education. So any advice to students who would like to go to multinational organizations such as the UN, in terms of preparing themselves, for the ever changing job market in the age of AI.

00:29:27:20 - 00:29:53:06
Jorge Martinez Navarrete
Yeah. I mean, I think that, first, one has to stay with the times and learn the new tools. Again, on the first part, learn them or use them as an end user. So, a finished product, incorporate that into the way that you– try it for yourself, for your calendar, for your emailing, for your mobile apps and so on.

00:29:53:08 - 00:30:30:18
Jorge Martinez Navarrete
And then, if you are into technology, data analytics, business improvement, consulting, you also need to know how the technology works and learn how to manage it and build it. Because this is super central to all corporate strategies right now. If you are not ahead of it, you are behind. And, I would also put it to you as a great opportunity as a new professional, for a young professional to position yourself in a competitive position.

00:30:30:20 - 00:30:52:06
Jorge Martinez Navarrete
These are the people who have been in the market for decades. You have the advantage. You have the opportunity to beat them and their game. You can say to a 10 year, or 20 year experienced, seasoned IT professional who has not done their homework or who are not keeping up to date that you actually know more than they do in this new field.

00:30:52:08 - 00:31:11:20
Jorge Martinez Navarrete
So you can be a leader in AI because you don't– Nobody knows this for more than, you know, a number of years. Of course, I'm simplifying things. People have been studying and preparing for AI for decades, but you can learn the new technology and be very competitive very quickly. So it's an opportunity.

00:31:11:22 - 00:31:32:17
Jie Ren
Yeah, it seems that it is a great, great new market. Right. Because it is news, uncertainty, a lot of uncertainty. And students can embrace the uncertainty to be ahead of the game, as you said, right. To be more advanced in terms of their AI knowledge than someone that is already in the industry for many, many years.

00:31:32:19 - 00:31:33:17
Jorge Martinez Navarrete
Sure.

00:31:33:19 - 00:31:38:13
Jie Ren
Thank you so, so much for your wonderful conversation Jorge. I learned a lot. Thank you.

00:31:38:15 - 00:31:39:00
Jorge Martinez Navarrete
My pleasure.