Evolution in Healthcare and Education Through Data and AI with Tom Andriola of UC Irvine

[00:00:00] Tom: The other thing that's happening is this concept of an unbundling. The traditional system basically says, you have to do all these things and you get the degree, that credential that is recognized by employers. If you do all but two classes, then all of a sudden I can get credit for 38 out of 40 because I can represent that I've mastered certain things. And in the employment world, Competencies and skills are what people hire for, especially earlier in the career.

‍

[00:00:24] Anthony: Welcome to Data Masters. I'm joined today by Tom Andreola, who's Vice Chancellor of Information, Technology, and Data, and Chief Digital Officer for the University of California at Irvine. He's a thought leader on how to use technology and data as a tool to innovate, invent, and disrupt. He spends his days working with the, UCI to create Data as a strategy initiatives and implement those strategies and enhance the new value propositions and manage the change that comes with those efforts.

‍

[00:01:26] Anthony: Welcome to Data Masters, Tom.

‍

[00:01:29] Tom: Anthony, thank you. Glad to be here today.

‍

[00:01:33] Anthony: So when I was thinking about, this conversation and this podcast, I'll be a little honest. I struggled, because you have a very wide mandate, in your role. as a chief digital officer because you operate at the intersection of two seemingly very different institutions, both under the University of California, Irvine.

‍

[00:01:54] Anthony: Obviously the one I think most people think of when I say UCI is the, world class academic institution. but there's also a significant healthcare provider under that umbrella. As well. So maybe to start, just to give a little context for folks, maybe talk a little bit about your roles and responsibilities, but also how that works across those two very different organizations.

‍

[00:02:18] Tom: Yeah, thank you for the question, because it is a unique role. You really won't find one like it, and you'll find things close to it, but it is custom designed, and it came out of an assessment that was done several years ago, and actually had a hand in kind of writing the job description as a vision for it.

‍

[00:02:35] Tom: Really world, where the world was going and the leadership here at the time, having a singular view of what they wanted UC Irvine to be. So in terms of, let's talk about, some of the things you just mentioned, right? So it is across two industry sectors, that work very different, that have differences, but also similarities.

‍

[00:02:54] Tom: and so. Healthcare is really for the most part, a for profit or, really a bottom line driven, type of organizational approach where an educational institution is, let's say more of a not for profit institutional mentality. so there's differences there.

‍

[00:03:08] Tom: There's certain differences in the speed and velocity at which. Things are moving, right? Universities shut down for periods of time when students go home, right? Or there are breaks. Hospitals, ERs stay open 24 7, 365, right? So you have those differences. But let's also talk about some of the similarities.

‍

[00:03:27] Tom: They both have this aspect of a very strong subject matter expert whose knowledge and experience is key to the value proposition. In one place we call it the clinician or doctor. The other place we call it the professor or instructor. And they're key to delivering on the mission of the organization.

‍

[00:03:44] Tom: And then you have student and patient have some similarities that they're heavily reliant on that subject matter expertise and the scaffolding around them. Another key similarity, and this leads to your question about the role, is they're both, information and knowledge centric industries.

‍

[00:04:01] Tom: Even though they don't may not think of themselves as they, if you think about what happens in an instructional environment or a research environment, it's driven around data to information to knowledge. and so when you think about healthcare, it's all about understanding through data, the patient.

‍

[00:04:18] Tom: And so the concept of, we're using more technology just generally in our society and how organizations work. Technology is really for me. A data generation engine. It's really what technology is. It basically takes something that we used to do strictly through analog, like you and I sitting in a room listening to each other talk.

‍

[00:04:39] Tom: It now turns it into this digital data stream that we can capture and review the transcript word for word.and we can look at these, images of you and I look at each other through a computer vision analysis, and it turns it into something digital that now makes it computable. Well, both industries are going through that transformation.

‍

[00:04:57] Tom: And so the concept was, understand where technology is taking both industries. Understand where the data was going to become critical from the standpoint of traditional analytics. And then how do you build data foundations underneath that? Because the future of A. I. really is predicated around having good data foundations so you can do something meaningful with the data through the use of these, current and new tools that are coming out.

‍

[00:05:24] Tom: And so they basically said, you know, we need someone who thinks at that level, who's thinking about not just today, but what are we going to look like as an institution that is educating students or caring for patients five years from now? And laying out those strategies and frameworks for us, and finding synergy across.

‍

[00:05:42] Tom: So we're not duplicating everything. And so that's how the role evolved and how I spend my time at the end of the day is really saying, where's technology taking us? What kind of data is it generating? What do we want to do with the data? How do we bring it together? Because, data silos exist in every organization and we're no different.

‍

[00:05:59] Tom: And then how do we start to think about. Analytics. Structured data. Unstructured data. What are we doing with generative A. I. tools? How does computer vision, play into a multi modal approach to thinking about analytics? So, you know, if someone's not thinking about that, or someone's only thinking about it 5 or 10 percent of their time, progress is slow.

‍

[00:06:18] Tom: But if someone has that as the core responsibility, then conversations get started quicker, and ultimately we get the strategies quicker.

‍

[00:06:27] Anthony: Yeah, and I really want to get to talking about the data, given this is Data Masters. I want to start a little bit more around almost the business of your business. and again, acknowledging that an academic institution, is less of a, for profit business, but thinking about the strategies of both organizations and your role.

‍

[00:06:48] Anthony: And as, as we talked about in the introduction is really tasked with Predicting the future and thinking a bit about the future, but not really about what's going to happen next month or even, I would think, next year, but really what's coming You know, five or even 10 years out, how the world's going to change, and how our relationship with data is going to be affected by that change and how we will affect the data we, to your point, the data we'll generate.

‍

[00:07:14] Anthony: So I thought I might share some. Somewhat controversial maybe points of view or ideas and get your view on them given your role. So maybe to start There's a lot of conversation today about the value of a four year college degree. I think it's still today Largely recognized as something very valuable.

‍

[00:07:36] Anthony: There's wage gaps between those with the a college degree and those without, but I'm curious on your point of view, on that value, is it possible that the four year degrees value will be diminished as more and more knowledge and information comes online more and more it's available for free through streaming platforms, And potentially that puts pressure on this wage differential between those with a college degree and those without, where either those without a college degree have the same knowledge and information as those with, or employers start recognizing that there may be less value there, they're not willing to pay as much.

‍

[00:08:17] Anthony: But I'm curious on how you think about that and how the school's thinking about that.

‍

[00:08:21] Tom: Yeah, so I think I would use the word these types of changes are inevitable and they're already on the way and I like to start here with something that we maybe don't think about is why did we used to go to live and go to school at the campus ? It's because that's where the knowledge was.in a time before the internet, if you wanted access to advanced.

‍

[00:08:44] Tom: knowledge around thermodynamics, because you wanted to be an engineer, you had to go to campus. You had to sit in the classroom. Well, after the internet came along and it really democratized access to at least information, then all of a sudden you could get access to that information through different mechanisms.

‍

[00:09:01] Tom: Maybe one led to you getting a piece of paper that said you passed a set of requirements and you were conferred a degree. But you could also get access to that in an informal way and have access to that knowledge and maybe master it better than the person who sat through the four years of class . So that was already starting to change and of course then technology caught up and said now we have different modes of how you can consume that knowledge.

‍

[00:09:21] Tom: You can sit in a classroom and you can do it online. you can be self paced or you could be, delivered through a traditional quarter or semester based system. So, these changes have already started to go and, of course, within the establishment, you have this argument over is the educational process better?

‍

[00:09:39] Tom: Is the pedagogy of teaching better? Are the outcomes better if someone who learns sitting in the classroom for 15 weeks versus doing a self paced online course that they may complete in eight? And that's been the debate inside. But we're now moving to a new era with A. I., which is, A. I. gives us the dilemma of can A.

‍

[00:09:59] Tom: I. act as that deep subject matter expert? If I can build a, someone who can deliver me the content, you know, a tool that can deliver me the content and spar with me in a one to one way, in a way that I could never do in a classroom where I have to sit with 75 other students, and if I can quote unquote pass the test, meet the requirements to get what would be an A, and master that, What does that mean to be able to say, quote unquote, I'm an educated person able to do job X?

‍

[00:10:30] Tom: The other thing that's happening is this concept of an unbundling. The traditional system basically says, you have to do all these things and you get the degree, that credential that is recognized by employers. If you do all but two classes, So I have to take, what is it, 40 classes.

‍

[00:10:44] Tom: I did 38. It's binary. It's zero or one. Well, if you start to unpack that into things like competencies and skill based learning, then all of a sudden I can get credit for 38 out of 40 because I can represent that I've mastered certain things. And in the employment world, Competencies and skills are what people hire for, especially earlier in the career.

‍

[00:11:05] Tom: And so these are the unpacking things that other institutions are starting to offer that, quite honestly, more and more institutions are saying, we need to follow this trend. The third trend I would say is the, just the acceleration, which is people, and this is, especially we're seeing it's in master's programs.

‍

[00:11:22] Tom: People can't afford to take the two years off. or the 18 month part time giving up, two weekends a month to get that. They really, what they're looking for is, I need short, concise, and I don't care if it's intense, learning of something that helps me get the next career step. So we have to offer packaging in a way that's different to meet the velocity at which the world and Specifically, the business world is changing.

‍

[00:11:48] Tom: So the question then comes back to, so what does an institution that has been very successful in the old model do in a world that is moving away from them very, very fast? And you start to see this with kind of the smaller liberal arts schools first. They're the ones that are impacted and closures going on and financial stress.

‍

[00:12:06] Tom: You know, institutions like ours right now aren't very impacted because we live in a state where the population is still growing. Uh, we have high reputation, you know, and also a research mission that kind of makes us a more attractive university. But we're trying to stay ahead of these changes and, for example, more than 20 percent of our lower level undergraduate courses now are offered in an online format.

‍

[00:12:29] Tom: They can take it in person or they can take it online. And so that optionality is what more and more learners are looking for, whether they're traditional learners, non traditional learners, or adult learners. And so we're trying to meet at the optionality level first, and then we'll move into different alternative credential models that essentially meet the need that people are looking for today.

‍

[00:12:51] Anthony: I love this idea that, As knowledge becomes more readily available, more digitized, it's therefore easier for people to access. To your point, you can now access it through other media besides, the in person experience. I'm sure the COVID experience accelerated all of that, and a little to your point, it also means that now that knowledge is available to models, to A.

‍

[00:13:18] Anthony: I. And I think this relates, to maybe the other side of your business, to the physician side and patient care and deliver and healthcare delivery. I have a lot of friends who are doctors. They often complain about patients who use Dr. Google. they'll come in and say, well, I Googled these symptoms and I think I have, fell in the blank.

‍

[00:13:40] Anthony: and what they're really saying there is, look, I have a specialized body of knowledge. Yes, Google is useful looking up a few facts, but you need me as the physician to filter those facts through my experience and my knowledge, to give you an answer. But to the extent that A.

‍

[00:13:56] Anthony: I. has been trained on all of the available medical literature, every single,piece of, literature that's ever been written, it could be the case that at 10 year timeframe, we have, models which are. better than, I was going to say as good as, but let's say better than the human doctor at, doing some of this diagnostic work.

‍

[00:14:16] Anthony: maybe they don't know if it's quite a good, bedside manner, but, but anyway, so I'm curious from your perspective, thinking about, as you train the next generation of, graduates, and then they think about going and working, at the healthcare provider, do, is that fair? Is that, or am I overstating with the way you see this going?

‍

[00:14:34] Tom: You're not overstating at all. The question is how fast and then every organization has to decide where you're going to be on this adoption curve, towards, new normal. So let me point out a couple of things. Since you have a lot of doctor's friends, next time you have dinner with them, ask them if they, A.

‍

[00:14:49] Tom: I., Uhm.So let's talk a little bit about the medical profession. and Jennifer and I have had a lot of fun broadcasting at Golden Horse, because that's one of our biggest goals and we like to share a lot of information with all our viewers. table If you call up Daniel Just to clarify, Dean, I'm in the medical school for six years, and I have a recommendation for you, first of all, everything is in the law.

‍

[00:15:04] Tom: wrong. We just don't know what 50%. Now, that's a scary thought as a patient, right? but there's a message in there, which is, look, we're going to give you a basis of education, but as our understanding of human biologies, our understanding of how our environment and our decisions impact our health, our understanding is going to change.

‍

[00:15:24] Tom: And so you have to commit to constantly reading and experimenting and relearning a better way, right? If you think about genetics. 30 years ago, what we knew and what we know today, if you haven't kept up with understanding how ge understanding and unpacking genetics drives certain disease categories, then you're behind the times and you're not doing right by your patients, right?

‍

[00:15:45] Tom: But already medicine, and medical care has known that it is a knowledge based game and that it's a moving target. So then the question is, how do you stay up to speed? When you start to put all this other knowledge out there, you start to change, and this is what some doctors, you know, kind of really bristle at.

‍

[00:16:04] Tom: Others are starting to become more acceptance of, or training doctors to be more acceptance of, is the patient who walks in with their own level of research. It means they're taking a level of agency of their own situation. And that shouldn't be a bad thing. It should be, a complimentary thing to create the right type of conversation between medical professional and a patient.

‍

[00:16:24] Tom: I'm actually part of a startup that basically is saying, look, I'm big into the wearable movement myself. I believe actually, what can my doctor tell about me once a year coming into their office into a very controlled environment and doing a set a battery of tests. that are a single point in time versus what wearables can tell me what happens over 365 days.

‍

[00:16:45] Tom: I was just telling some colleagues this morning, I have sleep stage data for the last four years of my life, every day. Good sleep, bad sleep, days I took naps, days I didn't get enough deep sleep. Isn't that more indicative of my true health nature? Then the doctor saying, So how do you sleep? You ever have problems sleeping?

‍

[00:17:05] Tom: What do they literally learn by that versus, Hey, I'm going to give you a four years worth of data. The reality is most doctors today didn't get a foundation in statistics to understand what do I do with that data? So one of the companies I'm helping launch is, It's helping doctors understand what to do with that data and integrate it into their thought process for how they see their patients.

‍

[00:17:23] Tom: So I think that's the other thing. Healthcare is realizing it's a data driven industry. It's looking to A. I. to figure out how do I deal with the massive amounts of information that we actually have around our patients that we generate that patients can generate. for themselves and now bring to the equation.

‍

[00:17:40] Tom: And how do we start to actually get different types of insights, maybe better insights through that to complement the things we might ask them in a traditional visit or in a traditional, stay that they may end up in the hospital. So what's a changing world. The question is how do you then move that back into the medical students who are starting this fall?

‍

[00:17:58] Tom: They're not going to start practicing. they'll start riding shotgun by their third year. They won't really start in their own practice until at least year five of their training. And of course, not, after their residency, they're really in practice. It's a long horizon.

‍

[00:18:13] Tom: If we try to shoot for where things are, How far off are we going to be just in what we've seen in the last 18 months since the introduction of ChatGPT? This is what makes it really challenging. It's like you have to continue to think about where's the puck going to be, and the longer that time frame, the harder it is to figure out the target you're shooting for.

‍

[00:18:31] Anthony: Yeah, it's a very much to that point. So the patients are coming into these doctor relationships with a lot more data and information, not whether to your point about your sleep experience could be tracking or exercise. It could be tracking. I have a digital scales. I have weight measurements going back years.

‍

[00:19:29] Anthony: And again, like, but to your point that doctor asks, well, what do you weigh? Okay. that's what I weigh today. It like doesn't tell you, has your weight changed? Like, it's my memory, that kind of thing. but also students are walking into the classroom with a lot more knowledge and information.

‍

[00:19:44] Anthony: And then to this question about, A. I. more generally, they're also walking in with a co pilot, whether it's, Dr. Google or Dr. ChatGPT, uh, or ChatGPT trained on medical materials, or even, you know, in a very literal sense, ChatGPT when it comes to academic papers. They're coming in with a co pilot who can, synthesize that information.

‍

[00:20:06] Anthony: And so, but I, what I hear you saying is, The job of the school and the healthcare provider is to add some value on top of that, to create a different kind of experience that synthesizes beyond simply what we can get out of the, data we carry around and maybe to your point, lean into it, embrace the fact that people are much more knowledgeable.

‍

[00:20:28] Anthony: So bringing these two ideas together for a second, your role at, University of California is to think about data strategy. so given this changing world, given that people have a lot more data, given that they are much more facile with that data, how are you thinking about the data strategy inside UCI,Is it just, dashboards and, uh, reports or hopefully we're doing more than that?

‍

[00:20:57] Tom: Yeah, so I think it's a couple of things I would point to. so one is I'd say the main strategies. started with kind of patient at the center in the healthcare enterprise and student at the center in terms of building that, that's who we're here for. And so let's build our data strategies around making sure that those two populations, those two core stakeholders or customers, I like the call them customers, have successful outcomes.

‍

[00:21:23] Tom: whatever defining assessment outcome is. and so then it's like, so, and now what is the universe of data that we're generating, that we've generated previously around this? Where is new technology in flight that's going to bring us new data or higher quality data? And how do we think about like aggregating it together, right?

‍

[00:21:42] Tom: And so you've got to add, you've got to pull it together, but then you have to give it Format and structure and governance around it, like build that foundation. And then we can start to create, that analytical layer on top of it that, that then helps people make data informed decisions.

‍

[00:21:58] Tom: That could be, if I take the student as an example. Okay, so that means the admissions people, the advisors, even the individual professors that they're working with. They now, we can now give them information for data informed decisions. In the classroom, that might be the context of, okay, this year's course that I'm teaching I have my same 50 students, but I noticed that everyone really did bad on this one, aspect.

‍

[00:22:24] Tom: of exam. Why? And then you go back into, let's talk about, let's look at the assignments we gave them. Let's look at the students engaged in the assignments. Did they watch the videos? How many of them were around the videos and actually watched the second, third, fourth time? Like that data now exists so that the professor can Self examine their approach to things.

‍

[00:22:43] Tom: And so like, so how do you then give them those tools? And so we've done a lot around what we call like the holistic and longitudinal understanding of the learner. Meaning from the time that they first express interest in coming to UC Irvine to the time they leave us and start into their career, We are, we have mapped out what is all the data that we're generating, how have we brought it all together, and building a personalized understanding, like a one to n equals one understanding of that student, and then personalizing how each person who's helping them get to those goals can work with them.

‍

[00:23:16] Tom: And so that's how we're doing it in the student, in the student realm. There's a similar story about how we're doing it in the patient realm. But there's a second part to your question that you asked, which is, how is that going to change? because, the analytics world is changing.

‍

[00:23:28] Tom: and I'll use this example. Regenerative A. I. Most of healthcare organizations, their analytical strategy was based around 5 percent data that was in structured form, and then figured out how to take all these semi and unstructured data Use natural language processing approaches and figure out how to pull that stuff out and put it into structured form so it could be part of the analytical framework.

‍

[00:23:51] Tom: That's what we're doing. Now all of a sudden you get this thing called generative A. I. tools, right? GPT. And now you can take all that unstructured data and now include it into your analytical framework and your way of thinking. So it went from 5 percent to 100 percent almost overnight on us . Now you have to think about how do I build an architecture to support.

‍

[00:24:13] Tom: Structured and unstructured continuously. How do I query it? How do you bring back the results around it? So it changed your whole strategy. In terms of at the data layer. But what we're only now starting to realize is not only is it changing kind of the data architecture layer, but at the analytical layer, the concept of building dashboards is going to go away.

‍

[00:24:33] Tom: If you've played with these tools now and they have to become more multi modal, if I could use that term, it's going to be a query based analytical world, which is, tell me, how many patients in my cohort that I'm responsible for do I have? on a GLP 1 drug, approved drug. How many patients, and and I'm going to say that in, a language, my core language as a prompting, and it's going to build me a graph and a narrative.

‍

[00:25:05] Tom: that is then going to be part of the report that I make. And it's not going to be specialists building dashboards for you and how do you want the filters to work. It's going to be every person is going to have the ability to speak to an A. I. agent, and it's going to generate the analytics.

‍

[00:25:19] Tom: And you go, no, no, no, no.What I really meant was, I wanted to understand the distribution by socio economic group or ethnic group within the population of patients that I'm responsible for. And then the engine goes and recalculates that. and so, this concept of it's not just a synthesizer, but it's actually a conversational colleague is where we're going, and I don't need the specialized skills of how to build dashboards or how to even operate dashboards.

‍

[00:25:47] Tom: I just need to tell it where I want it to do. We're trying to get our analytical frameworks ready to support that world, which is we're coming at it so fast, the question is who's going to be able to enable that first? And we actually showed some of our end users, like, what that world might look like.

‍

[00:26:03] Tom: Like, interact with the data, ask your question in the prompt window, look what's going to generate for you. And people are like, whoa, I didn't think we, we could do that with these tools. It's not perfect, but. Whoa, I didn't have to go talk to I. T. about building me a dashboard. I just asked this tool to build me a dashboard.

‍

[00:26:22] Tom: Have we bought it yet? Because I'd really like to start using this tool tomorrow. that's the world that's coming at us at such a fast pace that we're not ready for it yet. But the ones that are ready for it are going to gain advantage.

‍

[00:26:34] Anthony: So one of the common themes we've heard on this podcast is, data leaders, talking about creating data teams, which are really business partners. They understand the strategy of the organization. When they get asked a question like, how many of my patients are on a, a particular drug, They also know, to, push back and question, and meaning to think about the data context that they have, and then also challenge the questioner with what's the correct question to be asking about this data.

‍

[00:27:06] Anthony: and in that sense, I think, we really have an opportunity with these A. I. strategies where we're not just automating the dashboard production, but we're automating the data analyst because a good data analyst is a kind of a partner in crime. They're the one who thinks analytically. They think statistically, as someone whose undergrad was in math and statistics.

‍

[00:27:28] Anthony: I, like to think about these questions from that perspective. And if we can actually take that skill set put it very close to the data. it starts to really open opportunity about what data you collect, but also how you interpret results that come back. even some simple things like saying, well, there's, The results for this group are better than that group.

‍

[00:27:48] Anthony: Is it better enough? How much better? Is it better beyond one or two standard deviations? How do I measure that? And these are the kinds of things that a good analyst would push back on and say, yeah, this is different, but it's not different enough and those sorts of things. I don't know. Does that resonate?

‍

[00:28:03] Tom: It does. You know, I think, the generic data analyst has a value in the equation, right? And then if you have a data, a health data analyst or a health informaticist, right? They're more valuable because they understand the context of the data, right? They understand that the lab value is within a reasonable value or it's outside the norms.

‍

[00:28:20] Tom: That actually makes it more value in this equation that you're talking.I had a, let's call it a conversation about the future with a colleague of mine at another organization. And. We were talking about whether your data analysts, the people who are interfacing with your business, your business partners, if they need to become expert prompters, because is the key of the successful data analyst helping your business partner prompt in the most.

‍

[00:28:49] Tom: a. the answers to. And it was a really intriguing point, that he made to me, because I was like, I understand the power of good prompting in terms of, just doing it well, and being really good at it. And it's really interesting to think that we might be looking at a time now where we need to retool data teams Thank you very much.

‍

[00:29:12] Tom: into being the organization's expert prompters to help our business users get what they need faster. It's an intriguing question that I haven't wrestled completely to the ground myself, but this is like where the front edge of innovation is happening around data now.

‍

[00:29:29] Anthony: Yeah, to add a perspective to that, in some ways I think the degree to which we spend so much energy thinking about prompts is probably a failure of the system, more than, it shouldn't be the goal of the system. it should be the case that you can ask intelligent questions, it could push back.

‍

[00:29:46] Anthony: Go in to your point about entering into a dialogue, where, through intelligent interaction, it's guiding you to ask the right questions, get the correct answers. Now, all of this is predicated, on the underlying data. And so I want to, where we began the conversation and talk a little bit about data.

‍

[00:30:05] Anthony: And my limited experience with educational institutions is that they are absolutely the worst when it comes to data, in particular data silos, that they are often organized by department, by school, it could be split by a university and a healthcare provider, and they think about their world in a very insular way and as a result often create.

‍

[00:30:28] Anthony: really tall data silos that are very hard to break down. And obviously your role sits across all of those. And as we've talked about your strategy, for the institution, it's really thinking about data as a primary driver for that strategy. how do you think about breaking down these data silos, thinking about the quality of the underlying data, finding new insights when you look across, it must be a tremendous challenge.

‍

[00:30:53] Tom: It's definitely a challenge. I don't know if education is the worst, but it's It's challenging, for sure. You are not wrong in your statements about data silos. And this is one of the reasons why we had the foresight to put kind of data in the title, right? It was just like, someone has to be responsible for data, right?

‍

[00:31:11] Tom: Data silos means that you have data owners who want to hug their data, right? you have a lot of, the silos are really data huggers. And so the question is, it's like, Who tries to get the data huggers to release their data for broader institutional use? Well, that's kind of part of my role, right?

‍

[00:31:26] Tom: And so that means things like data governance. is something that I'm responsible for setting up and coordinating and trying to operationalize and, and have effective governance, versus ineffective governance, which is another thing that education institutions are really good at. so governance is something I work on.

‍

[00:31:42] Tom: And of course that we're talking really about humans, change management, and those types of things. And it is one of those things that kind of, my role, the level of my role allows me to be a strong convener of people, to. Push against the status quo of data silos and well, these are our students, in our school of X and to say, look, it's an institutional asset.

‍

[00:32:01] Tom: The data that's here is an institutional asset. You don't own it. there's aspects of it that the student owns their data and we need to give them agency of their data, right? But it's the institution's asset, not your school of business or your school of engineering to pick on the two of them.

‍

[00:32:14] Tom: And so my job is to break down those silos so that we build institutions. Institutional views partially so we can look at the differences in the cohorts that maybe are out of schools or out of departments or in STEM fields versus non STEM fields or students that are on Pell Grants versus, students that pay full fair.

‍

[00:32:30] Tom: So my job is to create an institutional data asset, which is an institutional strategy for how we manage, curate, and make available our data. So it absolutely is a challenge, but, I think the challenge is that every organization, it just might be a little bit. More intense and a little bit more challenging here because of the nature of local school and department autonomy.

‍

[00:32:50] Anthony: Yeah, so this very much speaks to, I have my, ongoing effort to, get people to talk about Dayton's law, for data, if you've heard of Conway's law says that the software that, Software company produces is a reflection of the organizational structure that they, or how they organize to generate that software.

‍

[00:33:12] Anthony: So Dayton's law is the same thing applied to data that the data that an organization collects will reflect the organizational structure. Of that they've chosen to organize by it. So in academic institutions, organizing by school, for example, creates this behavior of my students versus your students. It means that you tend to, in using your example, organize your or collect the data about students for a given school, and therefore the probability that you have students that cross schools that are not being found, goes up.

‍

[00:33:44] Tom: yeah, and I, you know, I would say like in like in our institution that behavior is stronger in the graduate student population than the undergraduate, where we have one process for all undergraduate students who come to us, you know, and this is where, you need someone who can identify that and raise a conversation because you can change.

‍

[00:34:03] Tom: And certainly you can, you may not have to change the organizational structure, but you can change the data flows in such a way that is no longer propagated in, by your structure. And I would say that we've had situations where my role has been the instigator and the driver of that.

‍

[00:34:19] Tom: as well as sometimes it comes back to, they lead us to see that we've enacted policies that just don't make sense in this kind of data informed world. And we have to have a conversation about how do we go change that policy at the institution because it's actually holding our students back from the outcomes that we want to see more of.

‍

[00:34:38] Tom: And so this is where data is really important. Sometimes you're using it in a very operational setting, and sometimes you're stepping back and saying, we have a structural barrier that we've created at the institution that we need to challenge about whether it is aligned with, the values that we have as an institution.

‍

[00:34:54] Tom: And there have been cases where, We've had that conversation. The reality has been, working with our faculty in one example, they're like, you're right. This doesn't make sense. There was a time we made that policy that we thought it made sense. Today, looking at it, it doesn't make sense. We can change the policy.

‍

[00:35:09] Tom: And that's a big thing in education, right? For the Senate, the Academic Senate to come back and say, We're willing to change a policy if there's something we've set in place. That's a big deal, right? But data drove the conversation for us, which is exactly one of the uses that data can play.

‍

[00:35:26] Anthony: Yeah, and I love that. This idea to anchoring it on the strategy. We want to create this outcome for the student. We can't create that outcome because the data is trapped in silos. Why is it trapped in silos? Because that's the policy we've created. The Dayton's law says we've organized it this way.

‍

[00:35:41] Anthony: And that way we're going to create this outcome. Great. Now let's change the policy and create the outcome for the, in this case, the student. By looking at the data in a fresh way, cutting across that silo, that's brilliant. I know, for folks that are listening, if these topics are interesting, Tom, I know you also have a podcast of your own called Digital Squared.

‍

[00:36:04] Anthony: so maybe if, for folks that are interested, maybe share a little bit about the podcast.

‍

[00:36:08] Tom: Yeah, thanks for the opportunity to plug. So, digital squared, life in an increasingly digital world is the tagline. And the whole idea is that our lives are really putting, being more and more put in a digital form. There was a time when the only way we could have done this podcast, you and I, is if I traveled to you or you traveled to me and we sat in a room from, and then we would, capture it.

‍

[00:36:30] Tom: Now we're able to use it over technology, which means we have. Files that we can compute against, we could correct things that were said wrong, take out the imperfections and all of that. Well, think about all the places where we now have, digital technologies enabling things. Computer vision is a great one, right?

‍

[00:36:46] Tom: Computer vision, looking at the retinal nerve. Finding diseases, you know, within it, right? championed by Google. So what we try to do in the podcast is bring innovators who are doing interesting things, taking advantage of digital movement and have them talk about how they got there, how they're pushing the boundaries of whatever their domain is and where do we think things are going next.

‍

[00:37:08] Tom: And, of course, A. I. is a huge part of that, but, we have to think about there's lots of other technologies. the concept of immersive experiences and digital twins are huge movements of technology that are changing industries even faster than A. I. in some respect. So, we bring a wide variety of technologies.

‍

[00:37:25] Tom: Thought leaders from different industries. They cover healthcare, they cover education, but we'll also bring people from, the venture community who are working with startups and what does an A. I. at the center startup look like today and how are they different from the companies you used to fund in the past?

‍

[00:37:40] Tom: So we bring a variety of speakers on there just to talk about how fast our world is changing and how exponential it really looks when you step back and look at how fast it's changing.

‍

[00:37:50] Anthony: Yeah, it was super cool. I'm sure, folks would love Really enjoyed that. I've really appreciated you sharing what you see Irvine's doing, how you're thinking about the overall strategy, where education and health care is going over the next five or 10 years, but more importantly, how to connect that back to data strategy and thinking about really how to enable and empower people with that data.

‍

[00:38:12] Anthony: to actually take advantage of that change and be ahead of it, as opposed to being impacted by it. So Tom, thanks a ton for joining us on Data Masters.

‍

[00:38:21] Tom: Anthony, thank you for the opportunity. Love the conversation.

‍

[00:38:23]

‍

‍