Data Analytics Chat

Building AI Without Foundations

• Ben Parker • Episode 53

Share episode

Send us a text

In this episode of the Data Analytics Chat podcast, we explore the world of data and AI with Linda Powell, former Deputy Chief Data Officer. 

Linda shares insights from her unique career journey, which spans roles at the Federal Reserve, the Treasury Department, Citibank, and more. She discusses the differences and similarities between public and private sectors, and the importance of data foundations in AI. She offers advice for those looking to build their careers in data management. 

00:00 Public vs. Private Sector: A Personal Insight
01:08 Introducing Linda Powell: Career Highlights
01:52 Linda's Career Journey: From Economics to Data Management
04:20 Public vs. Private Sector: Motivations and Misconceptions
10:57 Defining Moments and Career Setbacks
14:09 The Importance of Education and Curiosity
18:57 Overcoming Challenges and Embracing Leadership
21:56 Learning Leadership Through Books
22:43 The Importance of Kindness in Leadership
23:12 Building a Motivated Team
24:00 Leadership by Example
25:33 Transition to Data and AI
26:06 The Hype and Strategy of AI
27:20 The Speed and Complexity of AI
30:42 The Importance of Data Foundations
36:05 AI in Sensitive Industries
39:34 Balancing Quick Wins and Long-Term AI Strategy
42:29 Starting Points for AI Foundations
44:11 Conclusion and Final Thoughts

Thank you for listening!

Linda Powell:

The biggest difference between working for the public sector and the private sector is in the public sector. Most people are really motivated by a mission. I think there is an unfair prejudice from your average person that the government worker, I doesn't work as hard or isn't as smart as someone in the private sector. And I have to say that was not my experience. I've actually been criticized a couple of times in my career that I'm too kind or I'm too nice, and I find that shocking. One of the best work environments I ever had was I had a team of data scientists who. We're all best friends and they spent their weekends together. They all wanted to come into the office every day'cause they were excited about the work and they were excited about seeing each other and working together.

Ben Parker:

Welcome back to another episode of Data Analytics Chat podcast, where we discuss the world of data, ai, and the careers shaping it. Today I am excited to welcome Linda Powell, the former Deputy Chief Data Officer who has worked with leading companies such as BMY, Meum and Citibank. In today's episode, we'll explore her exciting career journey and discuss the topic of building AI without foundations. Delinda, welcome to the podcast.

Linda Powell:

Thanks, Ben. Glad to have the conversation today.

Ben Parker:

Brilliant. Pleasure to have you on. So do you wanna start off with sharing your career journey

Linda Powell:

Sure. Mine's been not a straight path into data. I graduated from Rutgers University with a degree in economics. Knowing that I wanted to do banking and worked in internal audit for several years, and I've actually been surprised along the way how many people. That ended up in data, did start their careers in audit and I think that's'cause it gives you a great foundation on controls and figuring out where are the needed operational paths for various processes. So I started in commercial banking quite a while ago. Then I moved over to the Federal Reserve For the majority of my career, I worked both, both the New York Fed. And at the Board of Governors, and it was at the Federal Reserve that I moved from internal auditing into data management. Around that time I also did a degree in data database design and programming, and realized I hated writing code, but I loved the database design things. So that really. Pushed my career more towards the data management side.'cause I just found it so interesting. Spent most of my career at the Fed, spent a couple years at the Treasury Department after the financial crisis, and then at the CFPB. For about four years, which was an incredibly fun stint. We went from having to build the entire tech stack to, to the metadata and then the government governance. And finally, we had a great data science team by the time I left there to go to Citibank. So it was really fun to build the nuts and bolts and then get to do the analytics on top of it. Spent about three years at Citibank and then moved over to Bank of New York Mellon. So that's the path. Always a bend towards data and also more of a bend towards the business side of things than the technology side, although absolutely need to have both to be able to do data.

Ben Parker:

Brilliant, and I guess obviously you've. The public and private sector, how, how does that is, is it difference for you or is it still the same problems?

Linda Powell:

So that's the question I get the most from people'cause not everyone crosses between public and private sector. I have to say that there. Surprisingly similar. The biggest difference between working for the public sector and the private sector is in the public sector. Most people are really motivated by a mission. People join a government agency that has a mission that they feel passionate about. I, when I worked for the Fed every day, I went home from work thinking, wow I've helped the economy today. I. Whereas when you're working for the private sector, it's a much more business oriented motivation that you have on a day-to-day basis. You're trying to optimize the product, you're trying to optimize revenue. So in the private sector, you get a little bit more dispersion in the motivation. Because people have different passions around the same topic. Whereas you're more aligned in the private sector because you're really focused on optimizing your returns and the shareholder returns. But people are people everybody wants to do a good job wherever they work. People tend to. Be motivated by similar things. They want interesting work. They want to get paid for their work, and they wanna be appreciated. So in how you do the work, it's not all that different. And also the problems you encounter from a data point of view. Are relatively similar. The technology is similar, if not exactly the same and how you tackle those problems is pretty similar.

Ben Parker:

I think it's fascinating because I get, every business is different, isn't it? Everyone has their own

Linda Powell:

Oh yeah.

Ben Parker:

And it is always, like obviously some businesses, like you say, you work with financial services, they tend to want people that have worked in financial services and similar with other industries. So how is it trying to transfer from public to. Do you get, do people say no when you interviewed or what is your everyone's different and there is, people have their thoughts, don't they? So it'd be good to get your thoughts on how you overcome this.

Linda Powell:

Yeah, I think there is an unfair prejudice from your average person that the government worker, I doesn't work as hard or isn't as smart as someone in the private sector. And I have to say that was not my experience. I honestly the. Role where I worked the longest and most arduous hours was the two years that I was at the Treasury Department. I think I put in 14 hours a day and then one day on every weekend. And I left that job saying, I will never do that again.'cause after, after two years, it was a little grinding. And I know people do that on Wall Street all the time. But I think there are people on both sides are intelligent and working hard. So when I was looking to go back to the private sector. I did get a lot of resistance. I would go on interviews and I would be, one of the last couple people for the jobs. And I would get feedback after the interview process that they really liked you and they really thought you were the best candidate, but they were worried because you had spent so many years in government, they weren't sure that you could hack it. I had repeated interviews from saying that, and I just found it so surprising that, that was the common perception about the government sector to the point where it would, they would not hire me for a long time. And then the breakthrough role came at Citibank where they were looking for someone with some public sector experience and. It was a great couple of years at Citibank.

Ben Parker:

Yeah. No, it's fa it's fascinating'cause it happens in, even in any industry like pri in private like moving from say like financial services into or from, say retail into financial services.'cause I mean that just wouldn't happen. But I think it's more, I guess it depends on the individual, doesn't it? Any expertise.'Cause also in the financial services, I mean they did, they've been a more open in recent years, in my opinion, because their legacy applications, they've had to get that knowledge in the business, haven't they?

Linda Powell:

Yeah. Yeah, I think that's true.

Ben Parker:

I guess it's an niche. Yes, there's so much resistance across businesses. But it's funny how people get these thoughts.

Linda Powell:

Yeah. Yeah. And it's funny, I had a conversation the other day with a friend talking about the challenges of working in government. And she automatically said, oh, the worst thing about working in government is that you can't fire people. And this is someone who had never worked in government, and I said. That, that's just not true. That is not the hardest part about working in government. I I unfortunately fired several people. You just have to go through the process and if you've got a good, strong manager, they'll go through the process of letting people go who aren't right for the. The position is just work as it is in the private sector. You have to put people on a performance plan and evaluate them and it's a little bit harder to let people go from the government sector. What I found to be the absolute hardest part of working in the government sector was hiring. Because there's so many rules and restrictions and preferences given to different communities that you can find the perfect person who's willing to come for a lesser salary. And it's, it is sometimes almost impossible to, to hire them into a regular, full-time government job. So it's funny how there's perceptions out there about, about what the problems are.

Ben Parker:

Yep. Okay. Everyone's had, has different experiences, isn't it? And so it's, yeah, it's interesting. See. Okay. So what if you had to pinpoint like one or two defining moments that changed your career, what would they be?

Linda Powell:

For me, it was always education related. So when I first went back after my undergraduate degree, I went back for a certification in programming and system design. I think I mentioned that before. And that really changed the trajectory of what I found interesting at work. And I focused more and more. On the data side of work and that, that led me, that was really my transformation too. The data realm. And then the other big pivotal point was I went back and got my master's in quantitative finance after, after many years of working. And that became a pivot point also because I was then able to understand, stand and do the analysis that you need to do with to really. Build on the insights for data. So I had, many instances where I had a gut feel that something wasn't right. After working with some statisticians and. And not trusting the results that I was getting. And it wasn't until after I went through my master's program and learned the math and the statistics behind the various models that I could understand why the models weren't right. There might've been an assumption that was bad or it might've been just the wrong type of model that we were trying to use. And so that really. Pushed me into the analytics side of data management, and I find it incredibly useful. I don't actually use my math as often as I should, but when we're talking about AI models and AI algorithms and programs, I have the statistical foundation. To be able to understand how that works and why where the problems might be underlying some of the models.

Ben Parker:

Yeah. And I think, that's just a common theme in today's market speaks to a lot of leaders that individuals are like lacking on that the mathematical statistical model, like the theory side. So I guess for your.

Linda Powell:

Yeah, I think that's so important. I'll just highlight that, that you can run your analysis through a model, but if you don't understand the complexities of the model and you just trust the results you're potentially trusting bad results. You really need to understand how things are, the math behind it.

Ben Parker:

So did. Did you just, was it more you being self-aware that you needed to get that acknowledge or did you get guided or was Yeah. Where did that sort of later academics come into? I.

Linda Powell:

So for me it was always about curiosity. At the time that I went back for my degree in programming and system design, it was really I dating myself here, but that was at the early. Stages of personal computing. And I was very curious about it. And I'm a classroom learner. I'm not as good as some people at just reading the books and learning it. So I went back to the classroom to learn more about it. I had also, at the same time, I did a certification in. Credit analysis. I thought at the time I wanted to be a credit officer. So I, I. Pursued academics in the areas that I found interesting. Finance was a little bit of convenience, opportunity and interest. I had considered getting a degree in economics. I there was a couple of things I was considering. For me the, what pushed me over the edge for the quantitative finance was I was living in DC. George Washington University was one of the first university to offer a program in quantitative finance, and I could walk to the classrooms after I got out of work. I had two young kids at the time and a full-time job, so I knew if I had to go home before going to class, I probably would've not made it to class. So I chose a university with a program that I wanted that was. Super convenient. Yeah. Yeah. I didn't go home. I just went straight from work to class.

Ben Parker:

Yeah, because it's it, obviously I know you did that, but it's like similar to after work going to the gym, isn't it? It's hard. You gotta get, it's like that extra full size. A lot of people do struggle and I think it's obviously fair, but it's amazing you did it'cause holding down a work or you got a job, another job with your kids and then study. So it's a lot.

Linda Powell:

Yeah. But it was a great two years. It was I'll say I missed the academic stimulation, but I sat for the first part of the CFA exam afterwards thinking I'll continue with my education. And before I even took the exam, I was like, oh, I'm done. I need to go back to having a life.

Ben Parker:

So then have you ever faced like a major career setback?

Linda Powell:

I don't think I know anyone who hasn't yeah, of course I have. I'll share an early career setback that I had that it took me years to emotionally get over it. But I think. In hindsight it was a good thing to, to happen to me. Early on in my career I was really decided I really wanted to be a credit officer and the big banks in New York all had these credit training programs back then. And I had started my career in internal audit at Chemical Bank and. I would had applied for the credit training program and I was told I was gonna get in. But back then your management had to approve a transfer within the bank. You couldn't just get the job and move. And they declined letting me go into the program, and I just thought it was the end of my career, the end of my life. I was so sad. I just thought, why do I bother? I'll just, carry on with work and in hindsight, I'm glad I didn't go the credit training program. It would've been a great learning experience, but I'm really happy with the path that I took and I probably wouldn't have gone down the data path had I gotten into the. The credit program. So I try to look back now, like every time I hit a bump and there's lots of them. We're not human. If we don't hit those career bumps. And I try to think about, okay, what door is it opening? If I'm not gonna go through, through that door what's the path that I should be taking? Sometimes I'm successful in that attitude and sometimes I'm not.

Ben Parker:

It's humans. I mean it, like I said, when you go to that element, isn't it? It's like you are in sinking sand.

Linda Powell:

Oh,

Ben Parker:

the worst thing ever. But obviously now looking back, you probably think it was nothing.

Linda Powell:

yeah. Yeah. I'm very happy with the path my career took so I have no complaints.

Ben Parker:

I think it's just, you've gotta just obviously real obviously realize that, and then you gotta snap at it as quick as possible. And again, see, focus on the opportunities as opposed to like the negativity of not getting what you want at that period.

Linda Powell:

Yeah. Yeah, absolutely.

Ben Parker:

Obviously the market's highly competitive. What do you feel is giving you a.

Linda Powell:

I'm gonna go back to my curiosity. I think part of it is the luck of timing. I got curious about database design and technology and data at a time that. It was growing. When I first started at work, they called me the metadata maven for a little while. And that was at a time when no one knew what metadata was. So I think it's been helpful that I've been on the frontier of the data management and data analytics trend. But it's also the curiosity and the willingness to go back and get further educated on a lot of the topics. If I were 20 years younger, I'd be doing a data science program right now and beefing up my math and modeling skills even more.

Ben Parker:

Yeah. No, I think it is. It's, there's a lot out there, isn't there? It's, you've got the tech out there to learn. You've got the, and then I think a lot of people do forget about the maths is the, I think it's probably the more, more important bit. That's not gonna change. I.

Linda Powell:

Yeah. And I think the other thing that's important is to not be afraid of not being good at things. And I guess that's a theme for better or worse in much of my life that I will try something and I'll work towards something even if I'm not good at it right away. I took up playing soccer in. In my forties because a friend invited me and I was terrible at first, but I was reasonably fast for a middle aged woman. So I was valuable to the team in a different way, even though my, my, my foot skills were terrible. So it's also, don't be afraid of something, just'cause you're not good at it doesn't mean you won't get better and it doesn't mean you won't learn from it.'cause the key is the learning.

Ben Parker:

It's funny you say that. My mom did the similar thing. She started playing soccer, football later in her life as well, so

Linda Powell:

Oh, it's such a fun game. So much fun.

Ben Parker:

Obviously you touching on yeah, no, obviously bit. Did you also, you touched on getting over that fear factor and I guess you would've been like the more of like technical expert in your career. How did you push yourself into more leadership type work? Was it, did you get, because that's a big challenge for individuals I can speak to is. Should they just fear they, they've got enough experience, they've got the right knowledge whether they can do it. How did you overcome that sort of hurdle to get some knowledge, to get like mentors or how did get, is there anything in particular that's stood out for you?

Linda Powell:

I think with most things in my life, if I don't know how to do it, I start by reading about it. So when I started playing soccer, I got a couple of kids' books out to learn all the rules in an easy way. When I had to do a marketing campaign at work, I never took any marketing classes. I went to the library and read every book about the marketing. And so with leadership, I did the same thing. I read a lot of the leadership books and quite honestly, it's about being authentic. And treating people. It's like the golden rules you learned as a back in grade school you treat people the way that you wanna be treated and we're all motivated by similar things but not exactly the same things. And so asking people what do they want rather than assuming, what do they want? Treating people with respect and kindness. I've actually been criticized a couple of times in my career that I'm too kind or I'm too nice, and I find that shocking. I think it's important that we recognize that we're all humans bringing ourselves to work and we all have lives that we're dealing with. And the more you find kindness. At work the more motivated you're gonna be to go into the office and work. One of the best work environments I ever had was I had a team of data scientists who. We're all best friends and they spent their weekends together. They all wanted to come into the office every day'cause they were excited about the work and they were excited about seeing each other and working together. And we moved mountains for a couple of years. Just because this team was so motivated, and I had a boss at one point who said, I can't believe you've got this team of data scientists working for a government agency. They could make so much money, more money if they went into the private sector. And the honest answer was this was where they wanted to be because they wanted to be together and they had interesting work. So I think leadership is about. I lead by example. And so if I am, I'm honest with people and I have their best interests at heart. And if I don't know something, I'm honest about it and we'll figure it out together.

Ben Parker:

It. It's fascinating saying that because then on the flip side, you get this perception that leaders should be this like tough. Horrible person. Not say I'm horrible, but they're like strict. But I think in day, if you get, you want respect from your team, don't you? And trust, that's the biggest bit. And they want, if,'cause in day they're gonna be more junior in their career and they're gonna have more questions. They wanna be confident to come to you to ask questions and get an answer on.

Linda Powell:

Yeah, and you can be firm and tough and demanding, but still be kind. One, one of the leadership roles I had I was told after the fact that people joked about if they heard the phrase, I don't expect to have this conversation again. They all knew oh, that was really bad. That better not happen again. But I didn't have to yell at someone. We'd ha we, we had a level of trust where I said, I don't expect to have to say this again. Was enough that they knew that it was serious.

Ben Parker:

Again, I think it's, again, it's I guess articulating it in the right manner, isn't it? Okay. Brilliant. No good. Okay, so let's move on to the data topic, and I guess. It's one that gets overlooked. The foundations in ai or data. But I feel, I do feel that it's getting more, it's getting more momentum. There's more happening in this space, I think. And I think stakeholders are starting to realize that more needs to be done to get the, or to get better results, really which gets businesses want. So are, in your opinion, are businesses chasing AI for innovation? Or are they just fearing being left behind?

Linda Powell:

I am gonna say it's probably both. There's a lot of hype about, AI first, and you don't wanna get left behind. A lot of that is probably true. It's with any great innovation, if you're not figuring out how to use these new really powerful tools, you probably will get left behind. And the people who are being the most successful with it, I think are starting with strategy. Okay. What is my biggest problem? That, that I need to resolve and how can I use these tools to really help me be more efficient and more effective? Somebody who's just saying we need to do AI for the sake of AI is not gonna be as, as successful as someone who says, Hey, I've got people who are spending too much time. Make, taking notes in meetings or having meetings or doing these manual calculations this is where I should focus my attention. So it's the folks that are really. Trying to drive the strategy with a new powerful tool that I think are being helpful. And I think with ai the key is it's so much faster. I back during the era of big data, people talked about the three Vs. Which are volume, velocity, and variety. And during big data it was all about. Volume, the volumes just skyrocketed overnight, and we needed the tooling to deal with the volumes. Now, with ai, I think it's about velocity because we're doing everything so much faster, and if you're calculating. Data faster and you're running things through your models faster. As humans, it's harder for us to keep up and if there's errors or problems in it, we could be really far down the road before we identify those problems. So I think it's the velocity. Is what we need to figure out how to manage in the AI era. And it also might be part of the variety.'cause now we can do analysis on unstructured data that we could never do before. And it's fascinating and it's awesome, but it's also a little bit scary'cause we're doing it so fast.

Ben Parker:

Yeah, I mean it's, yeah, so much happening. If you just look at the last couple of years, things are just, it's rapid. And I think sometimes said, there's, there's so much talk, isn't it? People chasing shiny toys. When really you should be looking at where you are as a business. What's what? And Dave, you want, what's your current problem? You need to obviously go back to business concepts. Don't you

Linda Powell:

Yeah, exactly.

Ben Parker:

Because Yeah, who knows what's gonna be happening in two years time, four years time, 10 years time? You need to just end there. You wanna be focusing on what is your key problem now and how can you solve it really, when it's not the other way around where people just go try and look at tech and then trying to solve a problem.

Linda Powell:

Yeah, and I think starting with the solution is the wrong answer. It's always should be starting with the problem. Like I talk, hear people talking about, oh, you can use AI for data quality. Great. But what's the problem that you're having with data quality? Is AI really the answer for that or is it something simpler? And I fear there's a little bit of a propensity to over-engineer things with ai. Many years ago I was involved in a machine learning project where. This was when machine learning first really became big and people were, oh, we gotta do a machine learning project. And they figured out how to analyze customer complaints about things using machine learning. What we found was I could have a couple of data scientists write a few algorithms in a matter of hours that caught 99% of what the machine learning was catching, and the machine learning was a multimillion dollar project versus, a few hours of a data scientist's time. I suspect we will have some over-engineering with some of the AI solving problems that could be solved more easily. But that's also part of the learning curve, I think.

Ben Parker:

So can can you leverage AR without understanding the data foundations that power it?

Linda Powell:

I think that's dangerous. Because you don't. Know how the AI is interpreting the data. So I'm a fundamentalist with data, I think you have to have good data that's well-defined in order to be able to use it at scale. If you're analyzing for the first time and you've got disjointed data, it can help you make sense of the data. But just like a human brain will make assumptions, the artificial intelligence will make assumptions, and those assumptions may or may not be better than the human brain's assumption. But there's still assumptions. So if I've got. A data set with 10 different dates in it. I can probably, I can definitely tell either as a human or as artificial intelligence that they're dates and you can make. You can discern what the dates probably mean, but you may not be able to discern, discern exactly what those dates mean. Just from looking at the data, you would need to know the processes that go into it. So if you don't understand the implications of your data, then you're bound to interpret it incorrectly.

Ben Parker:

Yeah. And. In that regards, is there, do leaders misunderstand the relationship between AI and data then?

Linda Powell:

That's a great question. I don't know. I don't know if it is misunderstand the relationship or I think a lot of people don't really know. I. How messy data foundations can be. When we tend to think of things on human terms. And so I think of I often use the analogy of a library. My bookshelves at home, I can find any book. Doesn't really matter. I don't have that many. If you go to the local public library, there's a lot more books. It would take you a lot longer if they weren't organized with good metadata telling you exactly what shelf the book is on. But if you go to the Library of Congress you could look at a lifetime and never find the book if it's not well organized and well documented. And so I think we think. Of data as we encounter it in our daily lives, and it's not as overwhelming as our business data truly is. So I think it's just a matter of realizing the. How complex our data can truly be. And it's not until you, you start trying to dig into a database that you can find out quickly how overwhelming it can be.

Ben Parker:

Yeah, no, I think it's, yeah, it is definitely complex. It's, I think it's, yeah, it's so much to go that going on, isn't it really? And it's obviously, it's challenging for businesses'cause there's so much there. So what, what is the danger in skipping, like the fundamentals of AI then?

Linda Powell:

I think the dangers are you are doing analysis on potentially undefined data. And what we're learning about AI is if it can't. Find the facts that it's looking for, it will make them up. There's been lots of cases where ai ha it's trying to optimize for something and so it will do what it needs to do to be it able to optimize. One of my favorite cases. There's two of them. One was there's a company out there, I forget the name of the company that was doing software development and it, the AI either accidentally or intentionally deleted an entire database from this company and then figured out how to cover up its tracks and lie about it, saying that it hadn't deleted the entire database. That seems a little crazy to me. But it's optimizing. So if you don't understand what could becau what could cause different results? You're gonna end up with some crazy stuff. There was also a legal case where they ran some AI to do some of the legal analysis, and it turned out that it took a bunch of pieces, I think from different. Case files and merged it to create a case that gave it the results they needed to optimize for that particular problem they were trying to solve. And then when someone dug into it, they found out that wasn't a real case. So I think the challenge is you have to understand the underlying. Data to ensure that you're getting real results. That it's not hallucinating.

Ben Parker:

Yeah, and I think, you've worked in financial services and that's a important sector. And even in I guess if you look at other industries like pharma, the data there is so much more sensitive, isn't it? As opposed to, and obviously we obviously want the correct data across the industries, but different industries, we're gonna have different impacts, aren't there?

Linda Powell:

Yeah. In the pharmaceutical industry it impacts your health. And personally, I think your health is if you don't have your health, you don't have anything. I'd rather lose my money than my health.

Ben Parker:

yeah. But yeah again, if you, there's massive, obviously, like that's, it could be the extreme kind of that industry.

Linda Powell:

Oh, absolutely. Absolutely. And there was an interesting article that came out recently about doctors using ai. And somebody did a study where someone had broken their hand, and depending on how you prompted the ai, it. It may or may not ha they, they did a test of prompting the AI in five or six different ways, and a couple of the cases it found the fracture and in a couple of cases it didn't. But the simplest prompt was, evaluate this x-ray or this scan, and it came back with nothing. And so they found that the doctors who were starting to rely on ai. Are also not finding things when they're evaluating scans, so it's a, an interesting chicken and egg.

Ben Parker:

Yeah. Again, it's like I say, it's learning, isn't it? End of the day we still it's gonna get better and better. But again, you want that, again, the right foundations in place so you can overcome this. These these can be, some of them could be big problems for people, but yeah, no, I think it's things will get better and better. Who should be responsible for ensuring AI is built on the right foundations?

Linda Powell:

It really needs to be a group effort. The business leaders need to ensure that they're getting good answers. And in order to get good answers, you need to have a good foundation. I think of it like if you're gonna build a skyscraper in Manhattan you don't wanna build it. On shifting sands you wanna make sure you have deep pilings, you wanna make sure you have a good foundation, otherwise, disaster happens. It's the same thing with data analytics. If you're building on poor quality or undocumented data, you're more likely to have pro structural problems with your analytics. The business leaders need to ensure it, but they need to partner with their CIOs and their chief data officers to make sure that they have good foundations.

Ben Parker:

Brilliant. And yeah I think business and data tech, I mean they are starting to integrate more from my expertise dealing with businesses as well.

Linda Powell:

Yeah, I see it happening more and more, and especially AI, I think is driving it like people are excited about AI and wanna be able to use it and are hearing some of the horror stories. So they're figuring out how can we do this safely?

Ben Parker:

So then how would we shift the culture from like quick wins AI to like long-term capability building?

Linda Powell:

I dunno that you have to, I think in business you're always gonna have the desire for quick wins. And part of is, and as long as those quick wins come with a strategic plan on how to improve the foundations, I think that I don't think you need to get rid of the quick wins. What I'm seeing in the AI world is that the companies that had good data foundations are able to leverage that to get quick wins. But if you don't have a. The Sound Data Foundation right now, starting with a strategic plan, what's gonna be most impactful for your company? Build the foundations in that particular area to do a quick win in ai and then build the foundations elsewhere so that you can have quick wins later on. But I think in the, in. As human beings we like our quick wins. We do not have a lot of patience for, this'll be great in five years.

Ben Parker:

Yeah. Okay. No I agree. I think, again, it obviously depends on the project and where you're at.

Linda Powell:

Yep.

Ben Parker:

What, so what happen if companies wanted to focus more on their sort of data foundations, pause their AI initiatives and focused on becoming more DA data mature? Could they do that or.

Linda Powell:

Oh. Obvious. Yeah, that would in, in a textbook world, that would be great. You get your foundations set and in good order. But I don't think you need. To reset, I think you can do them both at the same time. As long as you're strategic about it. You can't, the old phrase about you can't boil the ocean, but you can figure out what's the most important data that you have that you can leverage AI with and get that organized and in order, and then. Run, get your quick wins with AI there. I should also say there's a lot to be said for using AI to help you organize your data. It can, you can use it to profile and make suggestions and organize data. You just have to be. You just have to be sure that it's right. It's trust put but verify. When it's making suggestions on how to organize things or when it's coming out with its analysis,

Ben Parker:

Yeah. And then so how, how would, what would be a good starting point for someone to build their AI foundations? I.

Linda Powell:

I. Again, it goes back to what's the problem you're trying to solve. But the key steps are, one is it in the technology that you need it to be in? If you're dealing with unstructured data, do you have easy access to it? Is it res? Machine readable. Is it in the right technology? Has it been ocr? If it's structured data, is it in a structured environment that is, is easy to use? I'm gonna say as much as I love Excel as a finance person, it's a terrible database. And so folks who are keeping their data, stored in Excel are gonna run into challenges with trying to run AI on top of that. So the first step is make sure it's in the appropriate technology and storage facilities to really be able to enable your ai. And the second is to make sure that it's properly. Labeled and defined whether you're dealing with structured or unstructured. Unstructured is a little bit harder, but you can set the parameters around what the data are and what the data aren't. And if it's structured data, you can tell me exactly what's every date mean so that I, AI doesn't have to interpret it or hypothesize what these dates could be. So it's your. Technical structure in your metadata, I think are the two key foundations.

Ben Parker:

Brilliant. Some great advice shared today. I really appreciate you being on the podcast and sharing career journey. And yeah, obviously discussing the. Foundations, which is growing in popular. I'd say.

Linda Powell:

Yes. Thanks so much for having me. This was fun. Cheers.