AI With Purpose: How to Build Products that Matter

[00:00:00] Lauren Moores: If you hand me over an Excel spreadsheet or a PowerPoint, that just shows me the answers without the insights or the context, without context, well, you can’t do anything. That’s true. You need to be a good storyteller to be good at using these LLMs. Why? Because you need to provide context. You can’t just say, here’s my data set,

[00:00:21] Lauren Moores: tell me about it. Without giving more context. You can’t innovate with LLMs without giving it a reason for what you’re going after. Right. Similar to in terms of stories, I have been in situations in the past where we go out as our marketing, but we’re the best ’cause we’re the most accurate. Nobody cares about that.

[00:00:40] Lauren Moores: They wanna know what it does for them. Right. You need to be the currency. You need to be what everybody else is using.

[00:00:47] Gordon Wong: Welcome to Hard Problems, Smart Solutions, the Newfire Podcast where we explore the toughest challenges and the smartest solutions with industry leaders. I’m Gordon Wong Head of Data and AI at Newfire Global Partners, your host for this episode. In each episode, we bring you conversations with top innovators and decision makers tackling the biggest issues across industries.

[00:01:13] Gordon Wong: Whether you’re looking for insights to drive your own strategies, or learn from the best, you’re in the right place. Let’s get started.

[00:01:22] Gordon Wong: Hi everyone. Welcome to another episode of Hard Problems Smart Solutions. Today’s episode is a real treat for anyone passionate about data, innovation and building things that matter. I’m thrilled to be joined by none other than Lauren Moores, also known as the Data Queen. With over 30 years of experience as a startup data innovator, strategist, and advisor, Lauren has helped shape the way companies from early-stage disruptors to biotech pioneers use data to drive decisions,

[00:01:48] Gordon Wong: scale, impact and navigate complex problems. From leading data and platform strategy at companies like Montai Health and Ultimate Medicine to advising organizations like USA for UNHCR, Lauren brings rare blend of technical depth, strategic foresight and a gift for storytelling that makes data really make sense.

[00:02:07] Gordon Wong: She’s also a longtime educator at Harvard, a speaker and a writer, and someone who has made a career of asking bold questions and helping others build smarter answers. Lauren, thank you for joining us today on the podcast.

[00:02:18] Lauren Moores: Thank you, Gordon. That was a lovely introduction.

[00:02:21] Gordon Wong: My first question is, do you remember how we met?

[00:02:26] Lauren Moores: No, I don’t, I, I remember asking you earlier like, what, conference or where were you? It was probably like some meetup or I don’t know, where did we meet?

[00:02:35] Gordon Wong: That’s right. Well, it’s, there’s a little bit of … I bring it up because it’s a little bit of irony. I heard you on a podcast. And it resonated was what you said, you really focused on a business problem, you also focused on things like data quality. I love the title of the Data Queen. Right. And I, I reached, I just reached out, like, and I saw that you lived in, uh, the Boston Somerville area.

[00:02:52] Gordon Wong: Uh, I didn’t know that you were actually a neighbor at the time, and so I reached LinkedIn and you responded.

[00:02:58] Lauren Moores: Of course, I like fellow passionate data lovers.

[00:03:02] Gordon Wong: So, but you know, I mentioned, I love the terms of Data Queen, could you tell us where that title came from?

[00:03:07] Lauren Moores: Ugh, I love it. So my colleague in my group at Compete started calling me the Data Queen, Udi Dotan, who’s wonderful and, uh, I loved working with him. But he started calling me the Data Queen with all the work that we’re doing at Compete, which was competitive intelligence browser and uh, basically ad hoc analytics using Clickstream data.

[00:03:29] Lauren Moores: And I’ve embraced it ever since because yeah, I’m the Data Queen. I live and breathe it.

[00:03:36] Gordon Wong: That’s awesome. You know, and it’s, I had to point out that Queen isn’t just a fun nickname. It’s a reminder that you are a woman in the industry dominated by men. And where in order to be heard, women need to invest in a brand.

[00:03:50] Lauren Moores: Yeah. You know, there is also the Math Babe. I think I was Data Queen before the Math Babe came around, but I do and she’s amazing too, and is very well published. I just like it as a way to personify everything that I’ve ever done because I work across industries and because I’ve built systems from A to Z.

[00:04:11] Lauren Moores: So it’s just it’s hard to say, Hey, I am a X, you know, like I’m a machine-learning person, or I am a data management person, or I am a data product person. I’ve done ’em all. So Data Queen fits.

[00:04:24] Gordon Wong: Yeah.Yeah. And so, you know, what kind of technologies and opportunities and data and analytics are you, uh, most excited about today?

[00:04:34] Lauren Moores: Well, I live and breathe using the generative AI technology right now. And it’s exciting and very overwhelming at times because it’s similar to a couple years ago or years ago when we were looking at IoT and all the different devices coming online and all the different data coming, uh, available or supposedly coming available.

[00:04:59] Lauren Moores: In this case, it’s the generation of data, whether it’s imagery or word or content. And there’s so many different ways to do it. And so many players right now think about search years ago when we had so many players and who are we gonna use? And it we’re at that point. That’s what I’m excited about.

[00:05:17] Lauren Moores: Uh, I just don’t know how to keep up with it all.

[00:05:24] Gordon Wong: Yeah, I don’t think any of us do.

[00:05:26] Lauren Moores: Yeah, exactly.

[00:05:29] Gordon Wong: so I find, you know, this would be a bit, a little bit of a leading question, right. But I find some comfort, but also I find it helps ground me to remind myself that it’s, it’s the business problem that matters, right? And then these are all technologies that help us accomplish to delivery against the business problem, how do you feel about that statement?

[00:05:46] Lauren Moores: Absolutely agree with that. Uh, I think that one of the things that, uh, you guys sent me on just a bunch of possible questions, one of ’em was something about data strategy. You can’t have a good data strategy without understanding the problem that you’re trying to solve. And it doesn’t mean that you have to use the latest and greatest technology, right?

[00:06:08] Lauren Moores: It doesn’t mean that Gen AI is the, is the best thing to do. It doesn’t mean that you have to build out this comprehensive machine learning model. It doesn’t mean any of that. You really should start from the basics. You should start to understand what is the problem and what is the data I’m going to need.

[00:06:23] Lauren Moores: Now, let me just do a manual analysis of what this should look like and then figure it out from there.

[00:06:32] Gordon Wong: I hear a quote from Lauren here that’s like, before you figure out your data strategy, you should figure out your strategy. Is that something you would say,

[00:06:41] Lauren Moores: yeah. Well, no, we’re, we’re mean, no, it reminds me of, of, being asked by, whoever was leading the company at the time, are we in the cloud yet? Do we need to be in the cloud? What exactly are we trying to achieve? Is it so we can add it to our, investor profile so that VCs are going to think we’re doing something interesting?

[00:07:01] Lauren Moores: That’s no different than any of the AI today. D o you actually need it to get the job done? That doesn’t mean that you shouldn’t be paying attention to all the different technologies. It’s just that not every technology is needed in every instance.

[00:07:15] Gordon Wong: Totally agree. Totally agree. So let’s, let’s talk about some of the jobs to get done then. Right? You and I both have an interest in healthcare. You’ve been in, you’ve been doing practicing for quite some time. So what problems around healthtech, healthcare are you excited about being able to solve? And how are you, how are you going after this?

[00:07:33] Lauren Moores: Healthcare is so interesting. I feel like every time I get into a new industry I, I get into a more complicated data-related industry. And I thought agriculture data was hard. Healthcare data is much harder because we are so unique as individuals and think about where all your particular data is regarding your health.

[00:08:03] Lauren Moores: It’s all over the place, right? I’d still have manila folders with all my kids’ records in it, right?

[00:08:10] Gordon Wong: Mm-hmm.

[00:08:11] Lauren Moores: Does that do them any good? No. So healthcare has the same problem as every other industry in terms of where is the data, can I get to it, can I marry it? Can I use it in a way that it hasn’t been used before?

[00:08:27] Lauren Moores: These are things that go beyond whatever technology we’re trying to use at the time, right? So how might I use the technology that we have today differently that helps us advance healthcare? I’m not sure yet. Right? Because we know that Gen AI is really good at seeing patterns. We know that there are software being developed using that type of LLM and NLP advanced technology, right? But we still don’t know the right combination to get us to the magic answer. ’cause there isn’t a magic answer. So we still need to solve the siloed data. The LLMs don’t have that data, right? We don’t have that data. I don’t even have my own data. So how do we do that?

[00:09:31] Gordon Wong: Yeah, that’s a, that, that’s a question I actually think about quite often myself. And then we get, we get clients come to us and they saying, Hey, you know, we want to do X, Y, Z. Great. What does your data look like? Do you think your data supports this analysis? And they say, we have no idea. Can we just use the model?

[00:09:47] Gordon Wong: Uh, well let’s have a conversation here. You know, and then, and then there’s all this education about what a knowledge graph is and what raw materials are, and how do you do processing, how do you apply expertise and so on. And it’s and it reminds me of the physical world quite often. So in the past you’ve worked on diagnostics, drug discovery, microbial products.

[00:10:07] Gordon Wong: I can’t believe you’ve done all these things. Right? So what’s the next wow product you’re working on? If you can talk about it.

[00:10:15] Lauren Moores: I am currently working for a startup in the dementia space, and the wow product is, to me, it’s, I don’t consider, wow, it’s game-breaking, right? I guess you could use the same words, however, we’re not using anything but your ordinary process of drug development at this point. And that alone is hard. So, in terms of, actually being able to apply the latest and greatest, I don’t think that’s where we sit, but I think in the future there is a role for that in understanding how we might be able to expand things. But I’m still trying to figure that out and it’s actually research that we’re doing now on how might we expand our platform using latest and greatest technologies. Specifically LLMs, without giving up any IP.

[00:11:19] Gordon Wong: I find those speculative questions are easy to ask, very hard to get the answer to. If you don’t bring the right combination of expertise around technology, expertise around the business problem, and a expertise around being creative. Right. It feels like you would be, it’d be a good space for you to be in.

[00:11:39] Lauren Moores: I I definitely find that in healthcare it’s a little bit harder. I think if you focus specifically on maybe the patient it’s a little bit easier to figure out what to apply and when. When you’re talking more specifically on the drug development side, it’s a little bit harder, right? And you have AlphaFold, you have other things, but there’s so many things that go into figuring out what is the right combination of, chemistry and, and biology to affect the right amount of people.

[00:12:15] Lauren Moores: I don’t think that we have figured it out.

[00:12:21] Gordon Wong: You know, in previous conversations you’ve, uh, you’ve talked about the dark matter in, uh, in, uh, in food and human biology. All those things we don’t understand yet. So how do you build data strategies when there are so many unknowns?

[00:12:33] Lauren Moores: Yeah. Great point. Prior to focused on drug development, one of the companies we were looking at the nutritional value of food and how might you create more robust food-based products, right, focused on specific chemistry that would help you with your overall health. However, to start with, we only know a very small fraction of what our food is made up of. And the dark matter of food there is we don’t know everything. And I’m talking about the good food, quote unquote good for you versus the bad for you food. And knowing, having a better idea of what those elements might be and what those molecules are, would definitely change the way that we might develop drugs because there is a correlation between your health and your diet.

[00:13:36] Gordon Wong: So what kind of, um, tying into is this scenario you think that AI and the current technologies we’re looking at can help with?

[00:13:44] Lauren Moores: We need to do the screening first, right? So you need the data first. You need to be able to go through and screen every type of food there is and understand what makes sense. I mean, it’s almost like you have to figure out, here’s this huge funnel, right? What are we gonna focus on first? And it’s like this trust triangulation about, okay, we think this food might help these diseases, but we can’t eat a hundred of those a day in order to get to that level of that molecule that will help us.

[00:14:20] Lauren Moores: So what’s the alternative? You try to create it, you know, synthetically or find it naturally and put it in a different form. But you, if you don’t know what that food’s made up of, we need to do that. We need to discover that. It’s similar to the sequencing, the gene sequencing and other, that, changed the way that we think about data being used for drug development and other. This is similar. And I’m hoping that more people figure out that we need more research here, we need more examination of what’s there.

[00:14:53] Gordon Wong: You know, oversimplifying the, the periodic table comes to mind, which is my, essentially my favorite knowledge product of all time. But it, it sounds like you’re asking for, like, we need to be building a periodic table of food in the context of human and animal biology.

[00:15:08] Lauren Moores: It would be huge. I mean, it, you’re, you’re not, it’s not the periodic table, right? It’s not a limited set. It is, it’s almost like an, an infinite set of molecules, many of which we don’t understand. I mean, you think about it in healthcare, right?

[00:15:25] Gordon Wong: Yeah.

[00:15:26] Lauren Moores: The human body, we don’t understand how the human body works really.

[00:15:29] Lauren Moores: We don’t understand all the ramifications. We don’t understand the, the pathways. We don’t understand how, how everything interacts together. And, this variable food that comes from the environment also has all the different pieces to it, right? You’ve got what’s happening, where was it grown, what was, what chemicals were used or not used?

[00:15:47] Lauren Moores: What’s the soil? What’s everything else? Then, you, you introduce that and you have to, you don’t even know what it’s made up of chemically, right?

[00:15:56] Gordon Wong: Mm-hmm. Mm-hmm. Yeah.

[00:15:57] Lauren Moores: Just think about like, that’s, that’s a huge, and I’m gonna use knowledge graph differently, knowledge graph of connections across molecules, right? I don’t even know how to picture it.

[00:16:11] Gordon Wong: Yeah. I’m very, I’m very reductionist. And what, what I’m hearing is the emphasis on the need to build metadata dimensions, context, before you run out. Yes, okay. You can go out and collect a trillion events, right? All this quote unquote ruined data. But you, if you don’t have the knowledge about the dimensions and the metrics and so on and so on, they try to put that into some kind of meaning.

[00:16:33] Gordon Wong: It’s not, you can’t unlock the value of that data.

[00:16:37] Lauren Moores: No, but you also don’t wanna build, you don’t wanna build a huge thing, right? Because then you’re boiling the ocean. You know? I like just what you know back of the envelope. What would you do for every country? What, determine the people who need the most help in terms of health? Determine what is the predominant thing that they’re eating?

[00:16:58] Lauren Moores: What are they eating? And take that and figure out what it’s made out of. Right? If you think about it that way, then you’re narrowing it down to what matters most for the globe.

[00:17:10] Gordon Wong: Mm-hmm.

[00:17:11] Lauren Moores: And it might vary by region, it might vary by country, it might vary by state. But if we do that, and we can look at and say, Hey, we’ve identified a hundred different foods across the world that are very relevant to everyone’s health.

[00:17:30] Lauren Moores: And we understand, and we, LCMS them, right? And we all understand everything. All the molecules are in there and we understand what those molecules are. That’s a whole nother like 10 year journey. Probably not 10 years with a given technology, but could be. Then what can we do with those molecules?

[00:17:49] Gordon Wong: Yeah, I, and I love that you’re saying don’t just go out and build that if you don’t have a plan, because I, that’s one of the things I heard, too. Right, right. Because I, you

[00:17:58] Lauren Moores: It concerns you.

[00:17:59] Gordon Wong: Well, you know, you know, I’m learning from your experience, I’ve done the two year project w here the requirements are stale and then you don’t really deliver very much value.

[00:18:06] Lauren Moores: Exactly. Oh my God. Waterfall. Oh my goodness.

[00:18:09] Gordon Wong: I, that’s right. Oh my God. It’s a whole nother conversation. Right? We can, we

[00:18:13] Lauren Moores: It is. It is.

[00:18:15] Gordon Wong: Definitely go there. So, and so I’m a big fan of, um, of purposeful data collection, like within some context for solving some problem. And I think, and, and I know you have an interest in companies like Whoop and Levels and Oura, and, uh, in fact, one of my, and you know, one of my old team members is actually leading a research terminal team over, so I stay in touch with them. So how do you see that impacting, what’s the potential there for the future personal health check, personal health. What differences can we make?

[00:18:42] Lauren Moores: Huge. If we can harness the data I think that we are still figuring out this is the maybe second or third version of our original wearables that where, you know. A couple years ago, maybe 10, we were thinking, oh, wearables, they’re gonna help all the advertisers know what you are, what your mood is, or, whether they can advertise for you now.

[00:19:05] Lauren Moores: Yeah let’s think about it instead in terms of the healthcare. I have used Whoop, I’ve used Levels. I’ve used my Apple Watch, which I gave up, and now I’m using Oura. And essentially I use it personally. And, but ideally it would be something that is more connected overall with the other people that are part of my healthcare. That would be great. You know, I am a, I’m actually an investor in Levels and one of the reasons is just the pure idea of being able to tie CGM data with food. However, the issue there is we really don’t do a good job. We actually don’t at all do a good job of recounting the food that we eat, right?

[00:19:57] Lauren Moores: So anything which relies on food intake diaries or anything like that is, is prone to huge amount of noise. We, you can take pictures, but you don’t necessarily know how much you actually ate of it. You could take pictures, but you don’t really know what’s it made out of. You could have the same picture, one could be made of very healthy ingredients and one could be all processed.

[00:20:19] Lauren Moores: You don’t know. So we need to solve that piece before this data gets to be beyond where it is now. I think what Apple Watch and Oura and others are doing in terms of, allowing me to know how things affect what I’m doing at the moment is good. The algorithms are not as personal as they need to be. I certainly know my body better than they do. But it’s a start.

[00:20:51] Gordon Wong: My, my wife’s, uh, my wife does a lot of bicycling and her fitness watch is constantly telling her she’s in the red for her heart rate.

[00:20:58] Lauren Moores: Yes, yes. Me too.

[00:21:01] Gordon Wong: The thing is, her, her, her resty heart rate though is in the forties and she has incredibly high max heart rate, so the algorithm is constantly trying to give her bad advice and I’m really in some ways, I’m actually, even despite being in industry, I’m surprised that they were not more differentiated.

[00:21:13] Lauren Moores: I wonder if they were based off of, uh, single sex data. Right. You know, goes back to clinical trials and other things. We very, rarely get the right distribution of people to test and to gather data from so that we can have the right, or quote the best foundation.

[00:21:35] Gordon Wong: I would love to spin up a whole separate episode and just talk about this. If you recall, I built the warehouse team over at Fitbit, and there’s a, and there’s so many things we could talk about there, about how we used our data. We did save the company with the warehouse, but I think we only delivered a tiny fraction of the human value we could have.

[00:21:53] Gordon Wong: And I would, you know, so I’d wrestle about that. I would love to talk to you about that sometime. Like how do we get more value out, all this data, all this work we’re putting into it. But let’s talk about strategy, right? There’s all this AI buzz right now, right? And like you, I’ve got people coming to me all the time.

[00:22:08] Gordon Wong: Are we on AI yet? Are we in the cloud yet? So on and so on and so on.

[00:22:12] Lauren Moores: Yeah.

[00:22:13] Gordon Wong: how do you help people distinguish between, AI strategy and AI theater?

[00:22:20] Lauren Moores: I can, I can talk more about my work at Harvard Business Analytics, working with a lot of the students. ’cause we focused a lot on when to use AI and where. And the whole, the whole program is, is bringing people up to speed on using the latest technology. But very frequently these execs go back and their company says, Great,

[00:22:51] Lauren Moores: now go build some AI for us. Right? Like, what do you want me to do? I had one student come back and say, I’ve just been told to build out a whole machine learning program. Well, do you have data? And she said, no. Well, great. That’s where you start. And once you get the data, you’re not gonna run, build a huge machine learning portfolio.

[00:23:16] Lauren Moores: You’re gonna do some pretty basic EDA right exploratory data analysis, see what you might be able to do, and you can build a very simple regression model to prove out whether or not this makes sense. From there, you can start building the analytics and the insights that you’re being asked to do, but you have to start off with Excel.

[00:23:38] Lauren Moores: Like you have to start off with Excel or, or now. Now you don’t need to. Now you can just throw it into and say, Hey, code me up this Python program so that I can run it in my Jupiter notebook. Right? Uh, no problem. But it’s not building out a platform, right? You’re you have to crawl. You have to crawl, walk, and then run.

[00:23:58] Lauren Moores: So it, the AI theater to me is, Hey, and I’m gonna be very careful here with this nuance, Hey, we need to do that because everybody else is doing that, or it’s the latest trend and it’s gonna help us get more money. Right? To… So that, so the nuance here is Gen AI, you need to understand it. You need to be able to use it, you need to be able to manage it,

[00:24:27] Lauren Moores: if other people are using it. That doesn’t necessarily mean that you’re using it within your product creation or your service creation, right? So that’s the difference between like jumping on, like we have to have it and we have to make it part of our overall offering. Maybe not. Maybe you’re just gonna use it internally and you should, because you need to understand how it can be used effectively. Does that make sense? That nuance.

[00:24:55] Gordon Wong: It, it totally makes sense to me. It totally makes sense. You know, and I’m picturing, you know, I, I tend to use a lot of, uh, physical examples and I’m picturing that toolbox, which has your favorite tools in it, right? Do you, that you actually know how to use and you carefully select versus, you know, your average homeowner who just like, fills a box full of things they don’t know how to use. And then they have to store.

[00:25:18] Lauren Moores: Yes, exactly. Yeah. Or buy it just because everybody else has it.

[00:25:26] Gordon Wong: And so now we know in healthcare we have, uh, we have companies with tremendous amounts of data. We have these really big companies and they are doing the right thing in AI, they are trying to put AI to use and, and they will say it’s, they will, say the right thing. The right thing in terms of like, Hey, we gotta have a business problem solve.

[00:25:41] Gordon Wong: But, you know, yeah, things go wrong. So what’s your guidance to these companies who are trying to operationalize AI? What should they be looking out for? What are some tips on making maybe fewer mistakes?

[00:25:56] Lauren Moores: Oh, that’s a very broad question.

[00:25:58] Gordon Wong: It is. I know. Hey, this is this… You know, we don’t mess around here.

[00:26:03] Lauren Moores: No, there you go. I don’t, let me think about that, because essentially I don’t know how they’re using their AI, so it’s hard to tell them whether it’s right or not right. But let’s, let’s just be abstract. They have a lot of data. You possibly could use AI. Again, it comes down to what are you gonna use it for? Why are you going to use it? What does it benefit you in this particular instance to actually use it?

[00:26:30] Lauren Moores: Now, that doesn’t mean that you’re never going to, but you have to have the strategy of today I’m able to do X. Right? However, I need to get to Y, I need to actually build out a whole different system in order to get to Y before I can even start using AI the proper way. And I also have to be thinking about the fact that technology’s gonna change, right?

[00:26:56] Lauren Moores: So don’t build something that can’t be reversed or changed. Modularized. Modularized. I guess that’s what I usually think about when I’m building out a system, no matter what year, whatever technology it is, can we replace it? Can we replace the modules that are going to change?

[00:27:19] Lauren Moores: So for instance, you build with AWS. Can I back that out and use a different provider? You’re building with a particular uh, like lab software. How integrated am I with that software that if I have to replace it because it’s not performing or there’s a better replacement, how much is that gonna, what’s my opportunity cost there? Right.

[00:27:49] Gordon Wong: Mm-hmm.

[00:27:49] Lauren Moores: Always thinking what’s gonna happen six, six months from now if, if not longer, so that you can build a system that’s agile enough to handle those hiccups, right? Where you have to be like, oh, this isn’t working. Gotta change things. Now does Gen AI change that in terms of the way it works? It might. I’m not really sure on the, on the generative piece, because it’s so flexible

[00:28:18] Gordon Wong: Mm-hmm.

[00:28:18] Lauren Moores: it’s so random sometimes, and, and, and in terms of the creativity that you can do that, you could think of it just as another data input to the overall system. So your system still needs to be able to handle whatever’s coming out of it. It may or may not change. I think I’ve just gone off target there, but modularize.

[00:28:44] Gordon Wong: Well, that actually no, that fits in well is what I’m thinking about, right? Because we know, look, we know that healthcare is a massive space in the, the US alone. It’s multiple trillions of dollars worth spend for just Medicaid, Medicare, right? And I think it’s, I think last I checked it, except around 21, 22% of GDP. So we all wanna be agile, but then we have these big, the biggest problems of being tackled by some of the biggest companies. And they all, they have a tendency to, they’re very slow to turn right, and they have a need to plan multiple years ahead. If you have to make some bets in terms of efficiency in, in the use of AI, what would you bet on?

[00:29:20] Gordon Wong: And where would you reserve your agility? And I’m thinking again, in the context of these large companies that are trying to tackle these really big problems, you know, payers of millions of, uh, customers, for example.

[00:29:31] Lauren Moores: I think that you’d be surprised. So I worked for a, for a while as a contractor for a big healthcare company, and the focus was more on providing insights and analytics on the data there were.

[00:29:31] Lauren Moores: I think that you’d be surprised. So I worked for a, for a while as a contractor for a big healthcare company, and the focus was more on providing insights and analytics on the data there were.

[00:30:28] Lauren Moores: You know, we still haven’t solved our overall business intelligence layer. And I think honestly, that could possibly be solved with some of the more recent technology. But we don’t even know what that UI/UX looks like. If that makes sense in the sense, so you’ve got this whole new paradigm.

[00:30:50] Lauren Moores: People think about it as a, as a chat box, but it’s not, it’s more than that. Right? It it’s a system. How do you present the results from that type of system in a way that is not this legacy software UI/UX that we’ve always done? Because it’s a different paradigm. So it’s, yeah. I’m talking future now. I, I’m not answering your question on the AI.

[00:31:18] Lauren Moores: What’s the best thing to focus on? It’s, it comes down to the basics. Just get the basics right, because then you, then you can be fancy.

[00:31:26] Gordon Wong: Well, Lauren, you mentioned presentation and uh, and I remind, I mean, one of the things I always tell new analysts, and actually analysts have been around for a while, is that persuasion is one of your core skills. You’re generating advice. If you can’t persuade people, take your advice. What have you actually accomplished?

[00:31:46] Gordon Wong: So, I know you’re a storytelling expert. Can you, you know, give us some tips on how to become a better storyteller data? If you have, if you have stories about storytelling, I’d love to hear those as well.

[00:31:58] Lauren Moores: Ah, some of my greatest times is telling stories, uh, in terms of what you can do with data and being in that role for a while. For one of the companies I was with was great because you had a whole group that was building great machine learning, audience data, and you had all of our brands and our leadership and, one group wanted to talk about the science and the other didn’t care. So you basically have to marry the two. And it’s like, why do you, why should you care? Here’s why I care. And actually that’s something I usually ask. It’s what are we trying to solve? Why are we doing this right? It’s like, do I care? You haven’t proven to me that I should care about what you just did or what this data is.

[00:32:48] Lauren Moores: If you hand me over an Excel spreadsheet or a power, a PowerPoint, that just shows me the answers without the insights or the context without context, you, you can’t do anything. That’s true. It’s actually, you need to be a good storyteller to be good at using these LLMs. Why? Because you need to provide context. You can’t just say, here’s my data set. Tell me about it without giving more context. You can’t innovate with LLMs without giving it a reason for what you’re, what you’re going after. Right. Similar to, in terms of stories, I have been in situations in the past where we, we would go out as our marketing, but we’re the best ’cause we’re the most accurate. Nobody cares about that. They wanna know what it does for them. Right. You need to be the currency, you need to be what everybody else is using. You need to be really, really fun but accurate? No. So in terms of storytelling it’s making sure that you give your listener right, or reader the reason for why they want to continue to listen or to continue to read.

[00:34:12] Gordon Wong: So you, you’re speaking to the person, right? Yeah. Yeah. I can give you an example. It’s actually When I first started working with, uh, Antal at Ultimate Medicine, he was a student of mine and he was writing up his case study and was about the company that he was building. Or he did build and for the life of me, I couldn’t understand what he was trying to say.

[00:34:40] Lauren Moores: And he is tenacious and wonderful, and he wrote, rewrote it five times. But the core piece of it this is why this matters, and here’s the reasons why. It took probably two more years before I understood exactly what he was trying to do. Similar to pitch decks, we’ve done many pitch uh, competitions at, at Harvard and been involved with those and help, help mentor those.

[00:35:09] Lauren Moores: Many times you have people who are putting something together because they think it’s really cool, but they have no idea whether anybody really cares about it and they don’t know how to talk to it. Right. So it’s always like why do I care? If you can’t tell me why I care in two minutes, then I’m not gonna listen.

[00:35:25] Gordon Wong: Right. And if, and if you can’t, you’re frequently telling me that you didn’t bother to find out what I care about. Right. Right.

[00:35:34] Lauren Moores: Yeah. Or you just don’t realize what the market wants. Yeah.

[00:35:37] Gordon Wong: Yeah, yeah. Common question that I’m coming across from both my peers and also, um, new people in, new to industries, like what should I learn?

[00:35:46] Gordon Wong: What should I focus on, and what’s it gonna mean to be a data professional going forward?

[00:35:53] Lauren Moores: Well, those are two, two different questions in my mind, but let’s answer the first. What to learn. For me, I focus on the things that I need in my day-to-day, personal or work. Otherwise, I would go crazy trying to keep up with everything. And case in point, when, uh, Gen AI first came out, when we actually hit the Open AI, uh, version where it’s oh, you gotta, you gotta play with this, right?

[00:36:25] Lauren Moores: I was like, oh God. I just don’t, I just don’t, I just don’t have the capacity because there’s too many other things going on. And then it was, well, you should just get to know it so you can understand how you might use it. But it really wasn’t until I had a consulting project where it’s like, Hey, we really want to understand how we might be able to apply this.

[00:36:49] Lauren Moores: And, you know, we’d like you to figure out some logic for us. Okay, sure. So I learn what I feel is relevant for what I’m doing at the time. So what, how do you translate to somebody who is in a data role? I think it’s keeping your ears open. I, I, I have so many newsletters that come in, right.

[00:37:14] Lauren Moores: And I follow different people depending on what I’m focused on. I have changed jobs so many times in industry, so many times that you learn who is the right person to follow in that particular industry. Then I put it in the back burner. Like, okay, I know where I could go if I needed to find more about X, but right now I need to focus on Y.

[00:37:40] Lauren Moores: So then I, you know, so you’ve got this kind of subject matter data type things, and then you’ve got the technology in general. For instance, I learned Python real early on. But I don’t use it anymore. But I know how to, if I need to. I have done some machine learning, but I’m not a machine learning programmer. If I need to, I know what to look for, but I also know how to manage somebody who’s doing machine learning, right? So just you have to figure out what exactly do you need to know? ’cause you can’t know everything. Uh, you, it, it’s you. Some people can maybe, but, uh, but I have limited capacity, uh, for that front piece where that’s what I’m focused on.

[00:38:24] Lauren Moores: So it’s, it’s more about where are you and what are you trying to do, and where do you need to use it. It could be it’s, I want to transition to a different career. I want to have a different role than I have now. You have to ask yourself that to really understand what you would need to deep dive in.

[00:38:44] Gordon Wong: How does that, you know, and then this starts going in a slightly different direction, how does, how do you put that kind of advice into practice for an organization? Right? We were talking about individuals, but how does, how does an organization embrace them?

[00:38:57] Lauren Moores: Right. Yeah, I mean, when I’ve built out teams and, and sometimes I’m the only person on the team to start, right? It’s more about system thinking. What do you need in order to support the everyday decisions for the company? And what do we have in place? What needs to change? How we get there? What do I need to learn?

[00:39:24] Lauren Moores: Do I need to learn that software, understand its ins and outs, or hire somebody who understands those in and outs, right? What’s the ulti again, it comes down to what’s the ultimate goal? What are we, what are we trying to, to achieve? And it’s making sure that you’re not adding something just because you know the person from 10 years ago and they just came out with something new and it makes sense and hey, you can try it, right?

[00:39:52] Lauren Moores: I’m always willing to do a proof of concept. If it makes sense for where, what the objective I have, but I’m not gonna do it just because it sounds cool.

[00:40:02] Gordon Wong: Wait, don’t just run out. Learn that new programing language just because it sounds really cool.

[00:40:07] Lauren Moores: I did, I did learn Julia once, but I’d never used it, so.

[00:40:15] Gordon Wong: Yeah, that’s, I think that’s very similar to, again, I think about, when I think about the four basic, the pillars of skills for an analyst, I think you need to know something about technology, of course, you need to have some sensor math, you need some subject matter expertise though, right?

[00:40:28] Gordon Wong: You know, product finance and marketing. And then finally, and for me, the four, of course, the fourth flake is communication or persuasion. And the combination of those four dictates what kind of analyst you are.

[00:40:39] Lauren Moores: Mm-hmm.

[00:40:40] Gordon Wong: If you’re really, really heavy on the math side, maybe you’re more of a data scientist. If you’re really heavy on the tech side, may, may, maybe you’re more of an analytics engineer,

[00:40:47] Gordon Wong: right?

[00:40:49] Lauren Moores: Or

[00:40:49] Gordon Wong: you wanna get

[00:40:49] Lauren Moores: subject matter. Yeah,

[00:40:51] Gordon Wong: That’s right. But if you wanna talk to the executive team, you better know something about the customer journey. And that that feels like that gets neglected and that it’s been neglected, uh, all the way up and down the chain, right? Because we get asked to build data warehouse.

[00:41:07] Gordon Wong: And again, what problem trying to solve, you know., Or,, write me a model. What problem are you trying to solve?

[00:41:11] Lauren Moores: Yeah, yeah. It’s interesting. My, I’ll never forget a meeting I had early on and, uh, I went to the meeting not prepared, prepared in my mind. I was just gonna talk. But the CEO looked at me like, where are your slides? I was like, slides, what? And I, I learned like that. And then when I had people on my team and similar thing happened to one of my team members, uh, wasn’t prepared at all,

[00:41:43] Lauren Moores: and I said, This is how you do it. You need slides and here’s why. Right? And unfortunately, that means that you are creating the insights. You’re then translating them into slides that you can then show to leadership whenever you’re doing something, think about how you are going to tell leadership, right?

[00:42:04] Lauren Moores: How would you tell our CEO? Don’t, it’s not me. It’s not you. How would you tell our CEO?

[00:42:11] Gordon Wong: Mm, mm-hmm. Yeah. You’re go going to the problem, right? You’re going to the people who are making the decisions.

[00:42:17] Lauren Moores: Yeah, exactly.

[00:42:17] Gordon Wong: Yeah. That’s, yeah, that’s great advice. And I think I, I find my, I know personally I have to repeat it to everyone who’s new in their career to people who’ve been around for a while and to myself constantly. It’s so easy to fall with technology, especially, especially

[00:42:32] Lauren Moores: I agree with you.

[00:42:34] Gordon Wong: Yeah. You mentioned BI tools earlier and, uh, and the need to just be able to do descriptive analytics. I heard, I basically heard you talk about the analytics maturity model, right? It’s like, Hey, don’t jump into prediction if you can’t describe the past. Alright.

[00:42:48] Lauren Moores: Business intelligent tools. Oh,

[00:42:51] Gordon Wong: If a BI tool vendor was listening to this cast, you know, what, what’s your wishlist? What do you want those tools to do these days?

[00:42:59] Lauren Moores: My, my favorite BI tool was Looker. Because I’m a coder and I, I want the flexibility of being able to bring in the data a certain way and marry it. It allowed for better marrying. It was, it’s hard. It’s like snowboarding. You have to learn a harder curve, but once you learn, you’re good. Right? Um, however, it was limited because it didn’t have all the fancy features and beauty of Tableau.

[00:43:27] Lauren Moores: Uh, more recently I relearned Tableau from 20 years ago or whenever it first came out. And it was okay. There’s always limitations. I don’t know if there’s a be all, end all. I don’t even know what to ask for anymore because I’ve, I’ve moved beyond that honestly, in, in the sense that I don’t, I’m not playing with quantitative data as much.

[00:43:49] Lauren Moores: I’m playing with words. I’m playing with logic. I’m creating, uh, I’m creating behavioral analysis, I’m creating answers to what should we build next, right. Or what should we market next, or how do we market this better? And that’s a different modeling. So it requires a different way of presenting the information.

[00:44:16] Lauren Moores: It’s easier to present the quantitative information if you think about it, even though BI tools are limited. But when you have to start you describing or presenting all the information that you created during your Gen AI process and then converge it without losing the robustness of the, of the intricacies, it’s hard.

[00:44:42] Lauren Moores: The human can’t handle it. So that’s something that, that’s that UI/UX thing that I haven’t figured out, right? How do I and, and, and, and it’s also a, you have to fit in that content, that data into existing paradigms and existing expectations for a company. So that which is based off of traditional research or traditional, uh, data creation.

[00:45:06] Lauren Moores: So you we’re in this juxtaposition that I’m, I, I really want to see where does this go, right? Because I don’t yet know how to tell that story, and I wouldn’t know what to ask a BI tool for it. And I don’t think there’s one that has addressed this yet.

[00:45:25] Gordon Wong: Who are you telling this story to?

[00:45:28] Lauren Moores: Uh, our clients’ brands.

[00:45:30] Gordon Wong: Yeah. Yeah. It feels to me to, you’ve moved further left. Excuse me. Wrong direction. You know, I think of, I think of raw materials. You’ve moved further, right? You’ve moved even closer to the business problem. Right?

[00:45:42] Lauren Moores: Yes.

[00:45:43] Gordon Wong: You know, these, these descriptive curves. Some event over time is our attempt to describe reality. But of course it’s a very limited description, but it feels like you are trying to move into more complex, uh, concepts and more semantics.

[00:45:58] Lauren Moores: Yep. More semantics, much more advanced analytics, but advanced analytics in a way that we haven’t presented before.

[00:46:07] Gordon Wong: And you’re trying to, I’m assuming you’re trying to leave a paper trail, right? I mean, if you’re, as you, as you move into these more complex concepts are closer to the human behavior, you still need probability and reliability, right? So.

[00:46:18] Lauren Moores: Yes. Yes. And we do, we we have the quote unquote digital paper trail. Uh, what, what’s the logic? What’s the output? Why did that output in, create the context for the next question? It’s still hard to, unless you are the practitioner, right? I think… So, I liken it to you have a machine learning team. They’re using it, say they’re using it to understand healthcare characteristics or phenotypes or whatever. And they, as a practitioner, know the data backwards and forwards because they’ve modeled it. But you have to take that information and you have to present it in a way that people who don’t know the data as well can make decisions. It’s very similar.

[00:47:11] Gordon Wong: Do you see new jobs emerging, new roles? Right? Like what titles? What are the new titles? I guess not to be too reductionist.

[00:47:23] Lauren Moores: I don’t know. You know, we used to think that, oh, everybody’s gonna be a prompt engineer. Right. You don’t, you, you don’t need to be a prompt engineer if you, if you have set up, uh, an agentic system and I’m talking about semi agentic or you know, rules-based or whatever. You can, rules-based semi agentic, agentic,

[00:47:43] Lauren Moores: ’cause we’re not all the way agentic yet. I think there’s a role here for the more creative person, right? The person who is telling the stories. The person who is writing, who knows how to write. ’cause a lot of us don’t know how to write or, you know, uh, ’cause we’re not asked to. Uh, there is a role there in terms of translation because there’s a different translation skillset than the math person, although it’s very math.

[00:48:19] Lauren Moores: But, so it takes my math brain to create the right system of logic for Gen AI. But it takes the interpretation of output and it outputs not numbers right to under, and you have to understand it in a creative sense and you have to understand it in an influence sense and where, how it drives the ultimate decision.

[00:48:53] Lauren Moores: So it’s you’re maybe creating a whole new role area or functional area that different people are coming into and they’re not coming in from, uh, computer science.

[00:49:06] Gordon Wong: Maybe the, maybe the new titles are the old titles in some sense. Like, I mean, feel like you know the storyteller. Right. You know,

[00:49:14] Lauren Moores: Yeah.

[00:49:15] Gordon Wong: And, uh, might be the first original human profession in some sense. All right. Back to healthcare, right? So again, we talked about like, you know, you need to have a strategy, uh, how you’re gonna use AI. So what’s, what do you see as big misconceptions around AI in healthcare?

[00:49:32] Lauren Moores: It’s, it’s the same for any industry. It’s thinking that AI is the magic bullet that’ll solve everything, and that you just use AI and you’re gonna be able to create new things that we haven’t before. It, it, it’s so much more than that. It’s, it’s still comes down to the data foundation, you know, similar, whether you’re using quantitative AI or Gen AI.

[00:49:57] Lauren Moores: If you have bad data like Gen AI, you, there’s tons of bad data out there. So you need to know, you need to be able to understand that and how to use it. So, yeah, no, it comes down to the data and not being a magic bullet.

[00:50:10] Gordon Wong: Couldn’t agree more. That’s right. That’s it. Any advice, any, what’s one piece of advice you might have for our listeners as we wrap up here?

[00:50:19] Lauren Moores: Stay curious, ask why. Don’t just do it to do it. Honestly. just ask why, why is this important? Why do I care?

[00:50:29] Gordon Wong: Hmm. Yeah, that’s, that is actually probably the most important piece of advice. And do you have any questions for me?

[00:50:39] Lauren Moores: Well, tell me what you’re really excited about right now in technology?

[00:50:45] Gordon Wong: Yeah, yeah,

[00:50:45] Lauren Moores: Particularly in healthcare.

[00:50:47] Gordon Wong: Yeah. You know, really it’s, it’s funny, it sounds, I. It’s almost boring. And I think that’s the trap is that I’m really excited about just driving more efficiency. Because when we’re talking about, um, when we’re talking about thing, uh, programs like Medicaid, Medicare, we’re talking about serving people, you know, 20 million people who are disa disadvantaged or not getting proper healthcare and, or, or just anybody who was trying to navigate the system.

[00:51:11] Gordon Wong: Efficiency leads to better outcomes. Efficiency leads to better patient experiences. When you are not getting your questions answered, right, because you can’t get to a person or you, or the question goes to a wrong person. That’s a bad experience for a patient. They’re going through anxiety, they’re going through fear, they’re going through, just confusion that they shouldn’t have to. Or when your bills don’t get paid because the, there’s the wrong code on your claim, that leads to a worse patient outcome, right? And so I really do see it, and this is the echoes of this through my entire career, that, being efficient, right, can drive better outcomes. So, and then, and I think I see a lot of potential for AI to do that, right?

[00:51:50] Gordon Wong: In terms of, and then the other thing about efficiency is that you can plug into existing processes. You don’t have to boil the ocean, right? You can start at the ground level and make things 20% better, 30% better, you know, without changing all the processes. And, and then, and you, and you all, you don’t have to look three years ahead of time, right?

[00:52:08] Gordon Wong: You can sort of, Hey, how do I make this weekly process better? Which then it adds up. So, so again, it sounds a little bit boring in some sense, but gosh, hey, if we can spend $2.1 trillion instead, $2.4 trillion, that’s material.

[00:52:23] Lauren Moores: Yes. And I think that you’re absolutely right and that, that does come down to the system thinking, how do I improve the system so that we can work faster? Anytime I’ve been in any, uh, role, if I get asked to do the same thing two or three times, I build something for it. It, it might be a macro, it might be, uh, something that I built with, a very quick and easy software.

[00:52:23] Lauren Moores: Yes. And I think that you’re absolutely right and that, that does come down to the system thinking, how do I improve the system so that we can work faster? Anytime I’ve been in any, uh, role, if I get asked to do the same thing two or three times, I build something for it. It, it might be a macro, it might be, uh, something that I built with, a very quick and easy software.

[00:52:58] Gordon Wong: Right.

[00:52:58] Lauren Moores: Getting rid of manual processes that have, are more error prone, particularly exactly what you’re saying here, right, in terms of the patient or bill paying, oh, forget about it.

[00:53:12] Lauren Moores: Right? Or or bill billing. Billing, I should say. There’s some doctors I know, I know that that’s gonna be submitted incorrectly. However, I really like that doctor. I’m just gonna have to deal with the BS afterwards, right? We should be able to fix those things. But it does come down to the system, right?

[00:53:29] Lauren Moores: Because we’re all using different data and we, if we’re not training people properly and we don’t have the right it, we don’t make it easy, right? It’s gotta be easy. It doesn’t, it’s not, shouldn’t be 15 steps. You know, it reminds me in, in, in building out my logic systems. Now you can easily over build them because you can build anything, but you shouldn’t build it just to build it.

[00:53:54] Lauren Moores: It needs to, it’s, it’s no different than, and building elegant code, right? You don’t need 20 pages if you can do it in five. You just, so, so figure out where you can just get rid of that so you make it easier for everybody. Yeah. I’m totally in agreement with you.

[00:54:11] Gordon Wong: Well, there, there’s a famous Mark Twain quote, right about like, I was gonna write you a long, uh, a a short letter. I didn’t have time, so I wrote you a long letter.

[00:54:20] Lauren Moores: Yes.

[00:54:21] Gordon Wong: Lauren, thank you for sharing your incredible journey across data innovation, AI strategy, and healthcare leadership. Your perspective on building both solutions, whether in startups or biotech, dietary work. It reminds us that solving hard problems takes more than good data, right? It takes strategy, it takes empathy, it takes storytelling, and it takes, you know, it takes really caring.

[00:54:41] Gordon Wong: So thank you so much for that and all you’ve done.

[00:54:44] Lauren Moores: You’re welcome, Gordon. It’s been fun. I appreciate you and Newfire.

[00:54:48] Gordon Wong: Thank you. Thank you.

[00:54:51] Gordon Wong: Well, to all our listeners, I hope today’s conversation sparked new ideas about how we can lead with data and how we can build this with purpose and how we can stay curious, right, as Lauren said. Uh, even when the answers aren’t obvious. Thank you for joining us today on Hard Problems, Smart Solutions, the Newfire Podcast.

[00:55:08] Gordon Wong: See you next time.