Unlocking AI Potential in Organizations - Sergio Gutierrez-Montero, AbbVie

Khullani Abdullahi (00:01.728) Hi, and welcome to AI in Chicago, an Illinois-centered AI podcast where we feature residents of our state using AI and innovation to drive value in their organizations, conduct research, or invest in AI companies. I'm your host, Khullani Abdullahi. I'm the founder of Techné AI, a Chicago-based AI governance and strategy firm. I'm pleased to have Sergio Gutierrez Montero, an AI strategy and an Innovation leader at AbbVie. I'm practicing my Spanish, Sergio is an economist and a designer by way of education with degrees from the Barcelona School of Economics and the Institute of Design. Before his current role at AbbVie, he has worked at Allstate, McKinsey, and Dublin, among others. For those of you who don't know, AbbVie is a leading global biopharmaceutical company. Sergio Gutiérrez-Montero (00:34.419) Fantastic. Khullani Abdullahi (00:56.766) dedicated to discovering and delivering innovative medicines for some of the world's most serious health challenges. They have a particular focus on immunology, oncology, neuroscience, and eye care. So their breakthroughs are improving lives around the globe. And we are interviewing Sergio from AbbVie because the company is headquartered in Illinois. Sergio, thank you so much for joining me today. Sergio Gutiérrez-Montero (01:21.038) Thank you for having me. So yeah, Sergio Gutiérrez, currently director of AI for HR specifically within the corporate functions. It's an RIT role, it's a straight role within HR itself, which is kind of remarkable and I think is what first compelled me about the opportunity. It seems like very progressive that we have those capabilities inside the business functions themselves. As you said, I've joined the company six months ago. Khullani Abdullahi (01:41.674) Yes. Sergio Gutiérrez-Montero (01:50.026) I come from a at this point, kind of long consulting career. So previously I was with McKinsey and Company, what I did, mostly business building, in essence, helping clients, incumbents, imagine startups and new products and services they can launch into the market. So we would help them with anything from framing the opportunity, figuring out the strategy, the implementation plan. execute with them and we would help them launch. So that was fun times. Obviously, as AI capabilities have already progressed, many of those conversations become like, do we use artificial intelligence to better serve our customers or better target our customers? So that was interesting. Right before I joined Appian and left McKinsey, I was the long-term strategy lead for for Lily, which is McKinsey's consultant facing AI chatbot. In essence, we put decades of our knowledge into a little limb. as a consultant, as an analyst, you would go and ask, can you help me understand this topic or can you help me connect with the right experts at the firm? So very long, at this point, longish career, but happy to be at AVI. Remarkable a company that you know whose products change lives Khullani Abdullahi (03:21.802) Very cool. Saria, tell me a little bit about, so you have this very interesting educational background, but you're also very technical, very strategy oriented, and your career has followed a trajectory where you will remain relevant and can drive value for organizations, large and small, from strategy and execution standpoint. I'd love to hear and for you to share with our audience, how did you end up at this stage of your career? And maybe where did you first encounter AI in the work or personal settings? And when did you realize that it would be transformative in the world of work? Sergio Gutiérrez-Montero (04:07.436) That's a good question. Also kind of a big question. So I think to your first question, the way I found myself at my current role is because in consulting, it's almost like you are, in a sense, very aware and very flexible in adapting to changes in the marketplace. So AI became this big thing. We've been doing like... AI machine learning for the longest time as McKinsey, but it became this thing that all of a sudden everybody was paying attention to. So we needed to kind of almost figure out what our point of view for our clients should be. And as a result for that, you know, we did, I build up a business, you know, in the e-commerce space, can, I'm not allowed to discuss much, but you know, how do you target the right people, the right customers with the right. offers fraud prevention in terms of like when people return things versus not. So that it was already embedded like data science was embedded in some of many of our processes. And then once that, you know, the current AI tide came to be, we started experimenting with large language models to solve for very, very small but also very sizable problem spaces. Khullani Abdullahi (05:11.902) Wait. Khullani Abdullahi (05:19.424) Okay. Khullani Abdullahi (05:33.405) Okay. Sergio Gutiérrez-Montero (05:33.484) Again, I'm not a liberty to discuss, how you help organizations respond to help create documentation and help each to states about their offers for insurance in Medicaid, So again, it was just very tangible use cases. And I think that's kind of... Khullani Abdullahi (05:51.859) Right. Sergio Gutiérrez-Montero (05:56.991) That's what draw me to it is almost like not AI as in the abstract, but AI applied for in very specific ways, in very tangible ways to very specific business problems. Khullani Abdullahi (06:07.21) How do you stay aware of and up to date with your own knowledge and skills? Sergio Gutiérrez-Montero (06:13.691) It is head spinning. The rate of change is head spinning. I follow specific people like Ethan Mullig is like from the University of Pennsylvania is the person who I trust the most in the LLM space, which is he's a believer. Bending Evans, which is a former VC partner and tech analyst who's more of the skeptic side. So I combine both the optimist with the skeptic. Khullani Abdullahi (06:40.573) Yeah. Sergio Gutiérrez-Montero (06:43.231) Again, the point from Benedict Evans is, traditional machine learning has been around for a long time and companies still are not able to leverage that in meaningful ways for the most part. So what's gonna make us believe that AI, this latest wave of AI is gonna be any different. So even though I'm a director of AI strategy and implementation, which is a mouthful, I'm also Khullani Abdullahi (07:06.163) Okay. Khullani Abdullahi (07:11.103) Yes. Sergio Gutiérrez-Montero (07:13.932) I'm really aware of the challenges that unlocking the value from AI represent, what's going to take to actually unlock the value of the technology itself. And it all comes down to people. And I can speak to that a little bit later. Khullani Abdullahi (07:26.73) So, yeah. One thing I will say is, is Ethan Molyke the author of Co-Intelligence? Sergio Gutiérrez-Montero (07:36.147) I believe so, Khullani Abdullahi (07:37.088) Okay, because that is how I'm familiar with him. I also follow him on LinkedIn and Twitter. And I think you're right in terms of like well balanced, well reasoned analysis, he is a great tool. So I might ask you to jot down those names so that we can share them with everyone else so that they can also follow and read. Because I think what I hear a lot of executives say is I'm drowning in information, and I'm having a significant challenge. keeping up to date and not knowing what information I need to internalize and act upon and what information is not necessarily as valuable. So I think when you identify people that you find to be robust thinkers, clear communicators, it's helpful to share that out. So something that you just said right before we shifted gears, or I'd like to shift gears, is to this question of innovation. So every company, I imagine, or most companies, the executive teams are having this exact conversation. What is our AI strategy? How do we integrate AI into either our products and services and solutions and or how do we leverage AI to improve productivity and worker efficiency on our employees? And at the same time, employees are thinking, are these tools going to replace my job? Are there a danger? Sergio Gutiérrez-Montero (09:00.778) Mm-hmm. Khullani Abdullahi (09:03.166) to my partnering with the technical team, right? What happens if I sit down with Sergio and I tell him all of my thinking and my thought process and my workflow on a day to day, week to week and quarterly basis, and he just reverse engineers that and adds an agent or an LLM. You're an AI pragmatist. So when we think about the AI innovation value chain, how are you thinking about unlocking AI? and doing so in a way that doesn't make employees defensive, but makes them excited to embrace the technology. Because I think this is the conversation companies around America and the world are having. Sergio Gutiérrez-Montero (09:46.782) That is, I think that is the crux of my role, like understanding how to unlock the potential of AI. And I'm going to hope everybody comes down on it a bit, like substitute a job with all its nuances and all these like, you know, kind of peculiarities is just extremely difficult. And to the point... Khullani Abdullahi (10:04.05) Yeah. Yes. Khullani Abdullahi (10:11.935) Right. Sergio Gutiérrez-Montero (10:13.501) of like automating a job requires first understanding what that job is, like to the ground level. The only people who can actually explain what a job is and looks like and works like is the actual people doing the job. So it's not like there's going to come, there's going to be a wave of AI kind of taking everything over. It's just like adoption. It's going to be in the enterprise at least, it's going to be much more, much slower than people. Khullani Abdullahi (10:18.89) Yes. Sergio Gutiérrez-Montero (10:42.996) think it is. Because we don't have, for example, things like the infrastructure. Most companies have the data in different systems, in a format that is not of quality. So that is almost like, for starters, there is an infrastructure play that needs to happen that is just many companies are sometimes not aware of. And then second is Khullani Abdullahi (10:43.924) Okay. Okay. Khullani Abdullahi (11:05.642) Can you, before you move on, can you tell us and detail a little bit more what those infrastructure challenges are and what would be required to kind of resolve some of them? Sergio Gutiérrez-Montero (11:17.929) So in essence, like, I'm going to talk in, in broad strokes, just so not like explain what, what the situation at my company is. But it says that it's like systems of record that have like non-quality data. The systems are not synchronized. the systems are not put in the single, in a single data repository in a way that is like helps clean that data more easily and kind of leverage the data for. level 101, like advanced analytics, let alone like machine learning models, let alone like something more sophisticated. So I think there is a lot of piping that people are not aware of that needs to happen to enable like to unlock the value of AI. So we've grown accustomed to seeing Chagi PT and the experience in magical, but that is a very, I would say it's not like making that work. that technically work for a larger enterprise, the arc of that is much longer. And it requires a heavy investment in data infrastructure and data piping in a sense. Khullani Abdullahi (12:28.064) Excellent. So back to innovation unlock. So the first thing that you're identifying that is a potential bottleneck for companies is infrastructure. What else is required when companies are thinking about unlocking AI innovation and unlocking AI value chain for their organization? Sergio Gutiérrez-Montero (12:48.521) Yeah. So the problem is that they are focusing on AI versus focusing on the problem that needs to be solved. And it's just like, oh, it needs to be AI. It needs to be AI. Or how do we use AI? The problem is not how do we, the question shouldn't be how do we use AI. It's like the question should be what problems we are trying to solve for the business, for the business users, for our customers. It's almost like why are some of, we have now this new technology that gives us intelligence, infinite intelligence at scale. Khullani Abdullahi (13:17.961) Wait. Sergio Gutiérrez-Montero (13:18.17) Like, how do we want to use that? Shouldn't be as the primary question. The question is, what promise can this help us understand and emphasis on the promise? Like you need to be very keen, keenly aware of what the problems of your customers are or your employees are. So you can then say, we actually this thing that we couldn't solve in the past, we can now apply AI and solve it. I'm thinking about, you know, in the HR space, like talent acquisition. I mean, anyone who has been in the job market and I know this from, from friends, it feels like screaming into the void. is kind of this, like my bar talking to you, it's a black box. It's frustrating. It feels like impersonal. And those are things that the company companies who have a reputation to protect don't want to be perceived like perceived as. So that's a business problem on the side of the customer, which is in this case, a job candidate. Khullani Abdullahi (14:09.801) Yes. Sergio Gutiérrez-Montero (14:17.915) How do we make sure that experience interacting with our company is more pleasant? Can we, because we have intelligence, can we start instead of the old model of like posting a job requisition and having like 4,000 people apply, can we actually not post anything, say this is the job that we need to fill and actually scan the environment for those people and invite them to apply versus open the box for. Khullani Abdullahi (14:36.458) Right. Khullani Abdullahi (14:46.1) Just open it. Yeah. Sergio Gutiérrez-Montero (14:47.334) Exactly. So that's a flip of the model that is now feasible with AI. But again, the problem is not AI. The problem is like you need to have the imagination to imagine a better experience and identify the right problems. Khullani Abdullahi (14:50.89) Yes. Khullani Abdullahi (14:59.722) Yes. Yes. If you were to advise an executive team that is trying to develop their quote unquote AI strategy, would you recommend that they start with like clearly identified problems in specific business units and craft a proof of concept around the role that AI could play in solving that problem? Or would you say start with a broad-based, like 30,000-foot view AI strategy and kind of implement it organization-wide? Like in terms of getting to value quickly, which model have you found to be more effective? Sergio Gutiérrez-Montero (15:45.343) I think they were all like overarching infrastructure, like strategy is useful at the very corporate level. You need to have a straight, you need to have a strategy for partnerships. need to have a strategy for your tech stack. Those are, those are things that naturally belong to the more like study realm. But at the end of the day, the study only gets needs to be implemented. So it's almost how you use a top-down approach, which is like, what is our Khullani Abdullahi (15:51.968) Right. Khullani Abdullahi (16:03.517) Yes. Sergio Gutiérrez-Montero (16:13.617) or trade-offs and decisions around which companies we partner with. Is it open source or is it a proprietary space? Do we want to buy everything or do we want to build part of our strategy? Those things that belong are perfect, belong out there. You need to have an overarching framework. But at the end of the day, to implement, to unlock the value from that strategy, you need to actually be very granular. It needs to be bottom-up. You need to understand, you know... If we are talking about, you know, HR, which is my space, what are the things where we actually leaving money on the table in terms of like, have outdated processes that haven't been updated in the last 10 years. What is the market opportunity in terms of like engaging with candidates in a different way? You need to be very product driven, very bottomless up. Khullani Abdullahi (17:08.233) Right. Sergio Gutiérrez-Montero (17:09.392) So it's almost like it's top down for some things, but it's also bottom-top. You need to kind of identify the use cases that are going to be the raw materials or the core components of your strategy. So I'm actually going through that exercise about identifying pockets of value and problems to solve and kind of structuring that in a comprehensive strategy. Khullani Abdullahi (17:34.442) How are you, what's the playbook tactics or strategies you are using to identify high value use cases in your organization? Are you doing surveys? Are you embedding yourself within business units? Are you doing external market research? Like how does an organization optimally go about identifying the high value use cases that are important for AI investment? Sergio Gutiérrez-Montero (18:00.847) So I'm going to just change the study. The study had for, the design had, this is a design strategy problem, which is like the basis of the world. did a Dublin like help our company identify through design methods, ethnography, a qualitative research market scans, all those methods, like creating a comprehensive picture of the dose opportunities and then use the same methods to also paint a vision that is compelling. Khullani Abdullahi (18:09.522) Yeah. Sergio Gutiérrez-Montero (18:30.669) and everybody can anchor on. So it's almost like an innovation process, like traditional innovation process, the way we would use a doubling. I bring that with me. don't call myself an innovation strategist anymore, but that's kind of the core of my practice. You don't need to have a job title. bless you. Khullani Abdullahi (18:39.455) Yes. Khullani Abdullahi (18:47.688) Excuse me. Excuse me. My allergies are extraordinary today, so apologies. Sergio Gutiérrez-Montero (18:53.198) I'm sorry. You know, luckily, I don't have those allergies. I had it one year, but it was awful. Khullani Abdullahi (19:02.378) That's okay. So innovation design, you're using qualitative quantitative, Mac, market scans, surveys, feedback loops at the core. So traditional information and market development gathering strategies to inform high value AI use cases or prioritization. Okay. Sergio Gutiérrez-Montero (19:24.895) Exactly. Exactly. So the first, mean, I can talk about this in very general terms. The first thing that I did when I first joined Abyss just running ethnographic research with our colleagues, like what are the issues and business problems TAs is dealing with total rewards, talent management. So it's almost like, just like put your Khullani Abdullahi (19:40.222) Okay. Yeah. Sergio Gutiérrez-Montero (19:52.942) I mean, it's always okay to have some hypotheses and track them and put them aside and then do your like more like, I'm new here. I don't know what's going on. Sit down and listen for the business problems because that's going to give you the, good catalog of potential opportunities. And then you need to prioritize. Is this a pain for like deep pain for five people? And it costs like a million dollars to solve. Like maybe that's kind of their ROI is not quite there yet. Khullani Abdullahi (19:55.999) Yep. Khullani Abdullahi (20:22.144) They'll have to keep the pain point. It's only five people. Sergio Gutiérrez-Montero (20:25.763) Exactly. It's almost like things that are like, you know, they are not ideal, but it's just like the solve for those things is super expensive, even with AI. you know, those are less of a priority. Things that affect are very painful for a very large set of the population and they are somewhat doable, somewhat feasible to solve. Then those are the ones that become uh, become more of a priority. Um, and also, um, it's almost like there's like work management practices and design strategy practices that are like evergreen. They don't, they are like best practices, no matter what the technology at the time is. the end of the day, to unlock the value of a technology, you need to make, make that technology adoptable. If you build something that people don't want or need or cannot use. Khullani Abdullahi (21:05.94) Yes. Khullani Abdullahi (21:10.56) Yes. Yes. Yes. Khullani Abdullahi (21:18.132) Yes. Sergio Gutiérrez-Montero (21:23.726) You don't unlock anything. You spend money building technology that nobody's using, therefore it's not creating value. Khullani Abdullahi (21:27.872) So I'm going to share a statistic that I have read and heard often. And you being at McKinsey for a number of years in the past, I think is excellent for this question because there's this statistic in the consulting world that like 70 % of new innovation and like tech adoption, change management processes fail. And they fail in large part not because they fail 100 % but they just don't achieve significant enough user adoption in order for the objectives of the change management initiative or the implementation initiative to achieve full value. Sergio Gutiérrez-Montero (22:18.115) Mm-hmm. Khullani Abdullahi (22:18.152) Right. And so I imagine, especially with your background at McKinsey and within the context of AI, what should people in your role at other organizations think about when they are designing and implementing new solutions so that they don't fall into this large two thirds of projects that won't achieve sufficient traction be simply because employees refuse to use them, right. Or don't use them fully or don't use them. Sergio Gutiérrez-Montero (22:38.763) Yeah. Khullani Abdullahi (22:48.066) So how are you, you mentioned that just building the solution isn't sufficient. You have to really drive adoption and you have to drive adoption by ensuring it solves an actual need. And off camera earlier, you mentioned embedding it into their workflows. So what are the considerations that are associated with, we've identified a high value AI use case. We have a proof of concept, we're piloting it. Sergio Gutiérrez-Montero (23:04.674) Mm-hmm. Khullani Abdullahi (23:17.62) How are you closing the loop when it comes to adoption, employee engagement, employee training, and just engagement in general? Sergio Gutiérrez-Montero (23:23.649) Mm-hmm. Yeah. So I have a bit of a radical perspective here. you need to push people to adopt something versus them gravitating to your solution, you've missed the mark in, in framing, in framing the problem and framing the solution. Like I have this, this thing about, you know, I've heard it is like, people don't use it because they, we are not communicating enough. People are not using it because we need, they need to be trained. Khullani Abdullahi (23:31.24) Excellent. Khullani Abdullahi (23:43.392) So, yeah. Sergio Gutiérrez-Montero (23:56.611) We need to do change management. Like if you put something in their people, in the people, in people's environment that actually creates value for them, like delivers value for them, you don't need to do those things. Like it's almost like, do we need to train consumers to adopt a tool? No, if the tool is valuable and solves a problem, people will start talking about that and talk word of mouth that option will happen. Khullani Abdullahi (24:14.528) Yeah. Khullani Abdullahi (24:24.628) Wow, okay. Sergio Gutiérrez-Montero (24:25.214) And I think that's something that. Khullani Abdullahi (24:27.616) That is radical because there's entire businesses in the economy devoted to driving user adoption, user engagement, communication, the entire customer service function. Sergio Gutiérrez-Montero (24:41.804) Those things are helpful to accelerate, but if you've missed the mark and people don't want to use it. Khullani Abdullahi (24:48.254) never gonna happen no matter how much you communicate. Sergio Gutiérrez-Montero (24:51.41) Exactly. You can communicate and train people in the out. If the thing that you are asking them or forcing them sometimes to use doesn't solve the problem, you're like, there is no amount of change management you can do for that to work. It's just not going to happen. And I've seen them in every color in this space. It's just like solve a problem in an elegant, successful way. Khullani Abdullahi (25:00.84) Yeah. Yeah. Sergio Gutiérrez-Montero (25:20.459) For the most part, we'll need those things. Khullani Abdullahi (25:22.632) Let me ask this because my ideas, the questions that I ask are geared towards actionable insights. So I'm putting myself in the hat of somebody similar to you at a different organization that is less mature than where your thinking is, right? They're still finding their feet. As I'm designing the high value AI use cases, this leader... how would they know, like what feedback, what metrics, what KPIs would let them know that they're on the right track, right? Like, are you doing, what are you doing in the design and development phase that allows you to be comfortable that when you do launch, it is actually an elegant, robust, effective solution that will have high adoption, right? Because I think a lot of companies are just building. and then just handing it to their employees and then waiting for the value, waiting for the work efficiency. Sergio Gutiérrez-Montero (26:19.564) Exactly. It's like people are not using it. Why are they not using it? And the first thing they think people think about is like, they need to come. They don't know they exist. We need to increase our communication and make a campaign or adopt like train people because, those are the failure modes, like people going to these assumptions like automatically. Because they don't want to face. the harsh truth of you build something that nobody needs or you're or wanted or you solve the right problem. That's a very, so they try to kind of almost like play some theoreticals about, know, this needs some change management. know, nobody, mean, Chachi PT, what didn't launch like with campaigns, it's just people like. Khullani Abdullahi (26:46.644) You to. Yeah. Khullani Abdullahi (27:00.704) Right. Sergio Gutiérrez-Montero (27:10.727) was worried about the experience and started telling their friends. Khullani Abdullahi (27:14.238) Yeah, I think it was the fastest growing app in the history ever. It hit like 100 million users in 30 days. Yeah. Sergio Gutiérrez-Montero (27:17.597) Exactly. Sergio Gutiérrez-Montero (27:22.824) It's because it was remarkably slick for what I mean. Now we would see if we think about the experience that Chai GPT had like three years ago and compare it with the experience now that was like old school. it's magic. But you know, the point being, the point being you need to solve a problem. You need to provide moments of delight, just a classic product management mindset. Khullani Abdullahi (27:34.248) Yeah. Yeah. Khullani Abdullahi (27:50.048) So here's something that I think your answer brings to mind for me, which is how do you know who to hire and with what skills to do kind of, like if I wanted Sergio Montero at my organization, what am I looking for? Because you have AI strategy and implementation in your title now. but you did it before, right? And the vast majority of people don't have a work history of working in that role because the technology is so new. So what combination of skills, mindsets, backgrounds do you think companies should be looking for when they're looking for AI strategy leaders? Cause you're not going to find people who already have AI strategy in their title. It's a very small group of people. Cause it didn't exist 12 months ago. Sergio Gutiérrez-Montero (28:21.236) Mm-hmm. Sergio Gutiérrez-Montero (28:36.776) Because it existed. Yeah. Khullani Abdullahi (28:41.64) So what has made you successful in your role from a skill mindset, background level? Because I, and I want you to particularly talk about like you have core skills in product and design and innovation and consulting and research that have been incredibly important and relevant in the application of AI fields, right? but you have these core skills. if someone was trying to hire for you, what would you, what do you think they should be looking for? Like how would you, if you were expanding your team, what are some of those skills? Cause we have a lot of graduates in MBA programs and undergraduate that are coming into the work environment and people are trying to figure out how do I fill these roles? What do I look for? So what do you think is making people successful in AI strategy? Sergio Gutiérrez-Montero (29:25.396) Yeah. Sergio Gutiérrez-Montero (29:29.279) I think putting aside specific skills, it's all about have you built things that have been successful and you've been able to launch. You can philosophize about what the right strategy is, which is fantastic, and you need some sort of framework, organizing framework, so people don't go too far off of the path that as a company you've decided. But at the end of these are hard earned skills. Can you prototype? Khullani Abdullahi (29:38.804) Yeah. Khullani Abdullahi (29:52.885) Yes. Khullani Abdullahi (29:56.191) Yeah. Sergio Gutiérrez-Montero (29:59.077) identify sources of drivers of value. Can you kind of create a concept that you can present and shop around and get buying? Those are like, those is not a philosophical contract. like, it's just like, can you, with all its challenges throughout, can you actually get from having nothing to having a product in market? And when I say market is not necessarily a consumer market, it's like in a company for employees. Like all that good stuff. Because at the end of the day, things die in the vine because somebody doesn't align or they disagree with your thesis or things like that. So it's almost like you need to have the hard skills of like knowing how to do research, knowing how to do prototyping, knowing how to do financial modeling, knowing how to do a market scan. you have kind of instead of building something, you may buy something that is actually already a good fit. So all those... hard skills plus the ability to communicate a concept and get by and all that and just getting things through the finish line. And again, not to this part, it's like our MBA students, but that's something that nobody teaches you. need to do it and just like fail and have moments of crisis, which I have as... Khullani Abdullahi (31:14.878) You don't learn it in school. Sergio Gutiérrez-Montero (31:26.526) Somebody why so let me say to have this kind of tissue to prove like I've made I I've made mistakes They excite me Khullani Abdullahi (31:32.766) Yeah, you have the wounds. Yeah. I'm glad you shared that because what I think is there's a disconnect in the market, especially from a labor standpoint, where people, think you're probably the first guest I've had who's mentioned that traditional experience. is incredibly important and relevant, right? Where people aren't just looking for, are you competent technically in this field? And they don't have 15 years of multi-stakeholder management, right? Because you're not launching a product in any organization of significant size if you don't know how to get buy-in from different groups and committees and drive consensus, right? And so I think... Sergio Gutiérrez-Montero (32:16.634) Exactly. And sometimes, sometimes it's not about driving consensus. Sometimes it's about, you know, not shaping the environment so you don't even need consensus. It's like, you know, we're going do this. I'm able to convince the CEO so we're going to launch no matter what, which is one of the beauties of, of working for having worked for McKinsey is like you would have sponsorship of your initiatives at the highest level. So you would Khullani Abdullahi (32:26.9) That's it. Sergio Gutiérrez-Montero (32:43.602) go down and just like boots on the ground, just go and execute and launch. And everybody would fall like naturally would fall, which is something that is just a huge advantage. Khullani Abdullahi (32:53.076) But at AppVee though, I imagine you have to do a lot more multi-stakeholder committees and like AI strategy and correct me if I'm wrong, but AI strategy almost fundamentally crosses business units and business functions. Sergio Gutiérrez-Montero (33:09.458) Yeah. Khullani Abdullahi (33:10.24) right, organization wide. So do you find yourself doing a lot of meetings? Are you doing a lot of education? Like, how are you disseminating the AI strategy across the organization? How are you closing the feedback loop at all levels? Like, what are some of the skills and tools and frameworks that you're using? Sergio Gutiérrez-Montero (33:17.17) Yeah. Sergio Gutiérrez-Montero (33:32.894) I'm going to, I'm going to mention what I'm doing. I'm not saying this is the resolution or that I'm doing especially well, because it's, still like trying to kind of refine the approach, but things like, you know, at the end of the day, adoption is about people's business. Like you need to have adoption from the ground up. So I'm doing like, have an AI HR community of practice where I share like, you know, Hey, for time acquisition, these are some of the five ways you can. Khullani Abdullahi (33:48.83) Yeah. Right. Sergio Gutiérrez-Montero (34:01.224) These are five ways you can use AI. Or if you start structuring our processes in this way, we can then apply AI in these three ways. So there's kind of a lot of education and awareness that is happening. But again, what I'm finding myself is that people are not dummies. They are not afraid of AI, eager to try things. They just need some guidance. Khullani Abdullahi (34:26.89) Very cool. Sergio Gutiérrez-Montero (34:28.413) And that's almost like the vacuum that I'm trying to fill now is just like giving them almost not the AI 101 materials, but the 401 materials, because they are like, in my view, like the people who know the business the best and the functions the best are the ones that should be owners of the transformation. Like I told my best, like most of the day. So one experiment that I'm doing is like, can I work myself? Khullani Abdullahi (34:40.424) Right, right. Sergio Gutiérrez-Montero (34:58.343) Can I just substitute myself? Can I organize or orchestrate my processes in a way that are like automatable? Like, do I need to show up for every meeting? Do I need to do my own research? Can I create agents that actually automate many of my processes? And that's something that I'm doing on my own, but it's something that would be healthy for people to actually do. The prerequisite for that is, do we have the tools and the platforms to enable that? Khullani Abdullahi (35:05.28) Right. Khullani Abdullahi (35:25.31) Yeah, for them to experiment within their own role independently. Sergio Gutiérrez-Montero (35:30.199) Exactly. So one thing that I'm going to do next is like, hey, people are talking about agents. Let me show you how I build my agent for doing competitive research. you know, those are things that live outside of the walls of, of Abbeys, so, you most people at Abbe cannot use, but we are like, hopefully at some point getting some of those toys. So I want to prepare the... prepare the troops to be able to adopt and use those proactively. Khullani Abdullahi (36:02.356) Let me ask you this. Not only are you leading AI strategy and implementation in the HR function at AppV, but you are experimenting on yourself, on your own, within your role. And so I talked to a lot of, I just attended a luncheon yesterday where some executives mentioned not touching AI at all. for the course of their work and feeling a little distant from the technology. Do you think it's possible to continue to remain relevant and successful if you because executives in the past? This is my my thinking executives in the past could remain hands off from the execution and from understanding the technology that other people were managing. Is AI different? Do they need to get their hands into the tools to effectively lead their organization? or can they remain one, two, three levels removed from using AI in their work? There are still executives, there is attorneys at large law firms that won't use email, right? They're still dictating things, right? Is that possible in the emerging world or will those executives... Sergio Gutiérrez-Montero (37:04.763) Thank Sergio Gutiérrez-Montero (37:16.101) I mean, it's perfectly possible, but it's not going to go well. Like, how are you supposed to make decisions on something such a fundamental shift without you almost having even a rudimentary understanding of what the technology can do? You're going to make decisions based on random assumptions, something that you've seen in the news versus actually experience the thing yourself. So I mean. Khullani Abdullahi (37:21.098) Thanks. Khullani Abdullahi (37:32.83) Yeah. Khullani Abdullahi (37:42.752) Right. Sergio Gutiérrez-Montero (37:44.953) Yesterday is going to be always people who are going to say, this doesn't apply to me, or I don't need this for my job. I think knowing more rather than less is always helpful. Yeah, I can see that those people have been becoming irrelevant pretty quickly given the rate of change. Khullani Abdullahi (37:55.55) Yeah. Khullani Abdullahi (38:03.296) because of the rate of the change. Speaking of the rate of the change, I was doing a series of January, February, no, February, March and April where I was doing AI 101, AI 102, AI 103 for... a life science incubator in Chicago. And I ended up changing my trainings the night before, no matter how well I prepared, because new things had emerged. And I couldn't go four weeks without having to even update, right? And so I was constantly updating, how do you build a strategy? Because strategy and implementation, especially in large organizations, you're looking at like two year projects. But the technology is shifting every three to four months. The capabilities are emerging, are emergent. What does it mean to even develop a strategy in such a fast moving space? Have you seen that before in the last 15 years in the work area? This is different, right? Okay. Tell me. Sergio Gutiérrez-Montero (39:01.59) No, this is just like, this is different. But again, there is kind of one thing is like something gets launched. That's different from something gets adopted. Like some things that you're seeing are like not enter product rate or like our experiments, they drive adoption in the consumer market, but they're like not things that, that can drive yet value in the enterprise. And at the end of the day, the study is you need to use some principles. need to do quality data infrastructure and data quality. You need to have clarity on the use cases, what you're trying to solve for. And you need to be able to, one principle is like, don't get married to anyone, vendor technology, because at this point, something else is going to come and it's going to wipe you out. you know, I like to use this example of McDonald's wanting to do conversation commerce through their like, drive-throughs and using IBM for that. then IBM was using old school technology. Khullani Abdullahi (39:46.528) Yeah. Khullani Abdullahi (40:04.906) Mm-hmm. Sergio Gutiérrez-Montero (40:11.043) to natural language processing and so on. And the thing was not working. then there is like chat, you can talk to it. It's fantastic. And then that just like putting that, you know, putting a stop to that project. And that'd be not like very public failure, almost like don't get married. Just anchor on the vision. Don't get hung up on the details about we're going to use that vendor of this technology because that thing is going to change. So just. Khullani Abdullahi (40:19.071) Yeah. Khullani Abdullahi (40:27.464) Yes. Khullani Abdullahi (40:38.207) Yeah. Sergio Gutiérrez-Montero (40:39.225) things, avoid vendor lock-in, kind of technology lock-in, all those good things, like make things swappable. When you build your tech stack, make sure you don't hard code any one vendor or any one technology into your tech stack, for example, so you have the flexibility to adopt what's gonna come. Khullani Abdullahi (41:02.026) That's excellent because in traditional software development principles, we talk about modular builds and we talk about containerization, right? Like there's some traditional software development principles and I'm hearing you apply those principles into the AI strategy in general to ensure flexibility, agility, and the ability to interchange vendors or techs because both of those are going to change. Sergio Gutiérrez-Montero (41:12.185) Mm-hmm. Khullani Abdullahi (41:31.518) so that your problem vision continues to move forward, or the vision of the problem you're solving can continue to be solved, irrespective of the vendor or the technology. Okay, I think that's great. Sergio Gutiérrez-Montero (41:43.901) Exactly. It's another example of an evergreen best practice. The same way you go about framing the value of identifying viable problems to solve. That's existed for the longest time. The same thing about making technology modular based on, in this case, sometimes microservice architectures, getting that thinking of not building anything monolithically so you are able to swap things in and out. That didn't come with AI. That's best practice, evergreen best practice you can apply to the current environment. Khullani Abdullahi (42:21.33) Excellent. What do you wish you had known before you took on this AI implementation and strategy role at AppV? For somebody who's considering taking on a role like that, what do you think they should be reading? What skill do you think they should be learning? Should everyone be learning Python? Should everyone be able to train their own model? At what level of abstraction? should people be approaching their careers if they're committed to a career in AI strategy in a setting. Sergio Gutiérrez-Montero (42:54.936) I would say no enough. mean, I'm not a coder, but I know some Python. We're going to set up the infrastructure so anyone pretty much with some level of, you know, this level of understanding of the technology can build their own models and build their own apps. So it's almost like you don't need to kind of frame models from scratch. You don't need to be a Python expert. Khullani Abdullahi (43:00.565) Good. Khullani Abdullahi (43:18.569) Right. Sergio Gutiérrez-Montero (43:20.706) But at least if you need to build an agent that requires some Python to build a tool, having some sort of like understanding of what the lines of code like mean, you don't need to be an expert. mean, you need to know enough to not say, like to be able to direct others who are the experts and having a base level of understanding of those signals helps you with leading teams, with credibility, with. Khullani Abdullahi (43:25.556) Yes. Khullani Abdullahi (43:31.924) Yes. Sergio Gutiérrez-Montero (43:48.298) understanding with making decisions. if I didn't know, you know, the different flavors of models, the performance and understanding that over time, large language models, for, for example, are going to become a commodity. So that means I understand where the market is going from a technical perspective, but also I can read the implications from a study perspective, which means do not get married to any one vendor or technology provider because at the end of the day, everything is going to be pretty similar, like opt for things like open source, things like that. So you need to understand a little bit of the, of the technology, but don't get too much in the weeds because once you get too much in the weeds, you start losing the perspective. I need to know, you know, do I need to know that this, this model in this benchmark over, you know, past or these are the model that happens like every day. Khullani Abdullahi (44:45.652) Yes, it does. Yes. Sergio Gutiérrez-Montero (44:46.296) 53 times, you need to be able to understand the pattern, which is like, this is moving so fast and there's kind of an escalation. This is going to come down to models are going to become commodities. So you need to account for that in your tech stack strategy. Khullani Abdullahi (44:58.911) Yes. Khullani Abdullahi (45:03.754) So a lot of enterprise, speaking of like things moving fast, enterprise organizations are known for moving quite slow. and sometimes being risk averse. For innovative enterprise organizations, what are they doing to ensure that they're moving at the speed of AI or attempting to move at the speed of AI? Because I don't think anybody outside of the frontier AI companies is moving slower than the speed of AI. Only the frontier AI companies are moving at the speed of AI. And a corollary to that, do you think smaller companies have an advantage or do you think this is so capital intensive that enterprise organizations do maintain an advantage when it comes to the question of how do you move and build and add value at the speed of AI? Sergio Gutiérrez-Montero (45:57.608) Is this from the perspective of companies that may be a biopharma company that may be using AI or the AI native, like we build models type of companies? Khullani Abdullahi (46:08.586) So I think the AI frontier companies like OpenAI, Anthropic, Mistral, DeepSeek, and then Cohere in Canada, I think there's only five or six frontier AI companies. I think they're the only companies moving at the speed of AI. Everybody else is either a mid-market, small-market startup, or an enterprise organization leveraging AI. Sergio Gutiérrez-Montero (46:26.039) Mm-hmm. Khullani Abdullahi (46:34.014) What I'm trying to figure out is among those different size organizations, who do you think has the ability to move at the speed of AI well? Sergio Gutiérrez-Montero (46:42.87) Well, the thing is that the ability to train a model is becoming more widespread. back in the day, was just a pure research exercise where only the likes of Google, Meta, and Microsoft would have the capital to invest in computing power. But as the computing power decreases in cost, as the models, the methods, to train models become more efficient. Like one example you mentioned is DeepSec. They didn't use the latest and greatest like GPUs. They use like, you know, the second, third generation, but they found methods. Exactly. They use methods that don't require as much compute to train very, very powerful models. So as the field continues to move from Khullani Abdullahi (47:24.682) But they use the latest math, right? Khullani Abdullahi (47:33.215) Yeah. Sergio Gutiérrez-Montero (47:40.426) few research to more engineering based. In the engineering phase, we're going to start seeing ways to optimize for resources. And compute is not going to be, mean, compute is going to be very important. Don't get me wrong, but it's not going to be just five people in the world can do it. Again, it's more of a, it's more of a, what? Khullani Abdullahi (48:01.162) So do you there's a lot of disruption? Go ahead. Do you think that because more and more, so model development and model training will become democratized because computing costs are decreasing. Do you think that means enterprise organizations will be susceptible to more disruption in the age of AI than they were in the age of IoT or the rise of the internet? Sergio Gutiérrez-Montero (48:26.521) No, mean, unless you're going to become less susceptible to, I would say, to disruption. I mean, if you're talking about a biopharma, which is the type of company I work in, you're going to be able to access those technologies at a lower cost. you you're going to be able to use them. Therefore you're going to be less susceptible to disruption in that particular layer. I mean, It's actually interesting for the business model of a biopharma. It's of interesting where something that was very labor intensive and require a lot of human intelligence, like identifying a drug candidate, you can scale that to a huge extent. Therefore, your claim to fame or your claim to win in the market is sometimes... Khullani Abdullahi (49:08.852) Yes. Sergio Gutiérrez-Montero (49:23.274) those things are going to be less factors versus your ability to execute, your ability to commercialize, things like that. I think it's interesting. I haven't thought of this hard enough to understand this, but there is going to be, terms of adoption, more of a democratization. And because there is democratization, smaller biotechs are going to be able to maybe just identify compounds or... by a drags at a higher rate compared to or equal in our rate equal to those of beer companies. So that's a beautiful innovation just like lower the barriers to entry and make people compete on other basis. Khullani Abdullahi (50:06.056) Yeah. Khullani Abdullahi (50:09.961) Excellent. We're going to do a rapid fire, some Chicago cultural questions. Before we do that, I'd love for you to share how you think about measuring success. How people should, how should people think about any KPIs or metrics for knowing that they're on the right track? And then any last thoughts you have about what you see, like where do you, you and I had this conversation 12 months from now, what do you think would be different? or has been added to the market. And then I'd love to hear like any timelines, because some people have timelines where they'll say, all of this will be irrelevant in 10 years, no one will be employed. So I know you're an AI pragmatist, so I'd love to also hear about your timeline and any of those final thoughts before I ask you some cultural questions. Sergio Gutiérrez-Montero (50:56.339) Thank Sergio Gutiérrez-Montero (51:00.947) Okay, so the first one was KPIs people look at. Khullani Abdullahi (51:03.924) Yeah, how should they know that they're being successful when designing and implementing an AI strategy? Yeah, that's your metric. Sergio Gutiérrez-Montero (51:09.566) Are people using it? You you can theorize all you want. It's like, people using something that helps them in their job, helps them do something in a more efficient way? You know, are people using it? What is the actual value of people using it? Like, is that a yes no? Adoption, yes no? And then, you know, how much are they able to shorten their workflows or the time to value or? those sort of things. At the end of the day, it's not about, you know, does this model outperform these other models? Like, is this a model good enough that people actually use it and creates economic output in a way? Khullani Abdullahi (51:40.818) Excellent. Khullani Abdullahi (51:52.348) Excellent. And then on the Doomer scale, are you an optimist? Do you think that we will all still be employed in 10 years? Or do you think things will be markedly different in 10 years? And then over the next year, where do you think AI and AI strategy is moving? Sergio Gutiérrez-Montero (52:09.629) So I think everybody's going to have a slightly different job, if not a highly different job. But again, all that depends on how quickly can companies and enterprises adopt the technology, which is not a given. Big companies are very ossified. So anything new, adopting anything new, a much longer time than not. So yeah, we're going to all have jobs. They are going to look different. And hopefully we can... Khullani Abdullahi (52:26.496) Right. Sergio Gutiérrez-Montero (52:39.156) shape the solution together. Something that Reid Hoffman has this book, Super Agency, that talks about, you know, at the end of the day, it's going to come down to how we want to use the technology and it's going to be a slow motion. We're going to be able to shape how AI impact our jobs. And that's why I want to, you know, almost like put forth this idea of like optimism and agency, because at the end of the day, are Khullani Abdullahi (52:44.842) Mm-hmm. Sergio Gutiérrez-Montero (53:08.499) Everybody's gonna have a role in shaping what the solution looks like. Khullani Abdullahi (53:13.266) Excellent. That is such an optimistic final note to end on. So it wouldn't be Chicago if I didn't do, I actually am from, originally from Somalia and then grew up in Minnesota. So I've only been in Chicago for the last decade, but I know they take their pizza seriously. So are you a deep dish or thin crust person? Sergio Gutiérrez-Montero (53:28.936) Yeah. Sergio Gutiérrez-Montero (53:33.139) I'm more into paella because I'm from Spain. So I'm going to leave it at that. Sorry, that's a must not answer. Khullani Abdullahi (53:38.912) You're like, I don't like pizza. I'm gonna put this out there. They're gonna know you don't like pizza, Sergio. Paella is what you're gonna choose. Okay. Sergio Gutiérrez-Montero (53:46.117) No, I love pizza, it's like, okay, same crust, but you know, if I have to between pizza and paella, I will choose paella. Khullani Abdullahi (53:54.13) You'll choose playa. Okay, that's fair because I would choose anjera over pizza any day. Okay, do you have a favorite summer activity that is a must do for you in Chicago? Sergio Gutiérrez-Montero (54:06.317) I just ride my bike everywhere while the weather is good and lasts. Yeah. Khullani Abdullahi (54:09.245) You write your breakthrough act? Last? Yeah. Do you, so are there, so if you're from Barcelona, you must like soccer. Have you gone to see the Chicago fire or no? You have. Sergio Gutiérrez-Montero (54:21.927) Yes, yes, but I'm actually I'm actually from Madrid. I just happened to study in Barcelona. So, I've seen that she got a fire. Khullani Abdullahi (54:26.93) Eat in Barcelona. so you're a Madrid fan, sorry. And then, do you have a favorite place you like to grab drinks in Chicago? Sergio Gutiérrez-Montero (54:36.088) Ooh, I like Glebu-Sjon, I like Avec. Yeah, those are very solid places. Khullani Abdullahi (54:42.528) Fancy, okay. Shouldn't be surprised that you're fancy since you're from Madrid. Thank you Sergio so much for joining us, for sharing your thoughts. I think this is one of the most actionable podcasts. For those of you who are joining us for the first time, you can subscribe. at ChicagoPodcast.ai and I'll be sharing an article and some quotes from Sergio so you can kind of take these insights and actually implement them at your organizations. Talk to you guys soon, thanks. Sergio Gutiérrez-Montero (55:15.153) Thank you.