AI Governance Isn't Compliance. It's About Humans with Victoria Gamerman

Hello, everyone. Welcome. It is time for Catalog and Cocktails, your honest, no BS, non-salesy conversation about enterprise data management with tasty beverages in hand. I'm Tim, joined by Juan Cicada. Hey, Juan. Tim, it's always a pleasure to be in people's ears. We're talking about data and AI and everything. And I am super excited. Finally, after a couple of years, probably, Victoria, we have Victoria Gamerman, who's leaving corporate world right now. and she's starting Human Centered Transformation, RWD Insights. She has so much experience, hands-on leadership, data executive. Victoria, it is such a pleasure to finally have you on the podcast. Finally, how are you doing? Thank you so much. I'm so excited. I've got my favorite beverage here for this time of day anyway. So I'm excited to be here. Thanks for having me. So again, we're pre-recording right now. So it is 9 o'clock, 10 o'clock. Anyways, Tell and Toast, what are you drinking? What are you toasting for? and or what's your favorite cocktail mocktail of the moment right now? All right. So right now I've got a favorite cup that has the name of one of my favorite independent bookshops on it. I love reading. I'm an avid reader. I think there's so much we can go into around how books and knowledge get remixed. And today I've got a little bit of chamomile tea. But find me at a different hour and you'll probably catch me with a espresso martini in my hand. That's kind of my latest go to a little bit of spice and everything nice, which is just how I think we're going to go in this conversation today. All right. You guys love that. Tim, what about you? What's your latest? I am drinking my morning coffee, but a special drink I'll call out is actually my wife and I were trying out different like mocktail types of things. And so we we bought some higho cocktails, which are like non-alcoholic and they taste pretty good. So, yeah, that's my little drink of the moment. I bought this weekend Campari. I've been saying I want to get some and I'm like, so I'm going to start my Campari. because I've been doing my Negronis and Bolledeers. Now I just like, okay, I got Campari with this spirit. What are the options? What are you going to do? That's a great use of track PPT for me. There we go. All right. Let's kick it off. So much to discuss. Victoria, honest, no BS, AI governance. What's on your mind when we say AI governance? So I am really loving the term AI governance right now, and I'm trying to use it and sort of see how it goes and jives. I'm curious to hear what the audience has to say and what's on their mind around the topic. For me, I like it. It's an encompassing term, in my opinion, that really is the gateway, the key that's going to unlock for organizations, no matter how small or how big they are, to go from, we've talked about AI. We've now maybe tried a little bit here and there, but to really implement, to really find big wins, how do we get there? And to me, AI governance will be able to be the term that captures a variety of topics to help us with that implementation. So I'm excited about it. I think we can hit a lot of different things and I'm happy to kick it off. Awesome. What are all those topics underneath? So a couple of things. I want to say all those topics, but then also kind of in your experience, we've been talking about it. What are the POCs, the pilots, and what is missing to really unlock it to actually turn it into really production right now? You know, everybody's talking, everybody's doing POCs and stuff, but how do we really operationalize this? That's right. And so this is what I've, in my new space with RWD Insights, but also from the past conversations, I've mentored tons of startups through NYU Stern and List Frontiers Labs, and just sort of seen a variety of organizations, you know, in the healthcare space. And so that's the space that I'm in. I'm in the healthcare life sciences ecosystems world. I see a lot of exactly little pop ups, little POCs. And we tend to do this when we see sort of new shiny objects. But having also a background in statistics, biostatistics, which is highly regulated, and then moving to data governance throughout my career and really thinking about, you know, how do we appropriately share data, especially when it comes to the U.S. ecosystem? I just last week was actually reading, and I think more and more will come out, around one of the first FDA warnings towards a company that was using AI in the healthcare. So a manufacturer. And so, you know, you're starting to now see, okay, well, you know, it's exciting and cool, but there are consequences. So I think about many topics, and you see here I'm not talking about digital transformation, which has been my corporate title for a while. and definitely a theme I still believe in. But now I'm starting to see the pivot that really to do that transformation, we have to focus on humans, because humans really impact a lot of the things we talk about. So I think about four things. I think about no surprises here, right? People, process, technology, data. Okay. And so when I think about all of those things, all those things really have a really big human person element in them, right? So you talk about people, of course, they're sort of the culture of the organization. What are the priorities? What are the intentions? And how do we go about doing things intentionally, rather than maybe what I sometimes have seen across a variety of organizations, mushrooms popping up, lots of little ideas, boom, boom, boom, right? They're everywhere. Everyone's doing a little bit of something, but that has so many challenges if we focus on people, right? And the culture, you can have a leadership that's really excited about AI. And so everyone's like trying a little bit of different things, but then there's no standards. There's no consistency in how even with your approved internal tools, you're talking to them. And so you get different information or, you know, if you're also on the side using one of your own personal tools and all of a sudden now it's conflating things, how do you, you know, understand what it's doing and how do you check it? And how do you still use your brain and how do you avoid AI slot, right? There's like a lot of human people components there. Process, also I would say, right, a lot of process is human. And I actually think that that's a big area that if you get right, you kind of evolve the whole ecosystem because people will follow a process, hopefully, right? They might, you know, but if you just have a process and you change one step of it to make it more AI informed, and again, so much to talk about there, you will then not really, you know, you can make something a little faster, but in clinical development, in bringing medicine to patients and people, right, one day is great, but 10 days are better, right, to get the medicine to them faster. And so I really believe we have to reimagine an entire process, right? And process, we need that process still, does it, you know, end to end, how do we, so you have to really understand, which again, is really human centric because a lot of organizations have very invisible processes. A lot of people who are really smart, have a lot of stuff in their heads and they add magic to the process. And so there's a lot of these invisible things. So process is the second thing. Technology. If you think that AI is a technology issue, you got to watch more of the Honest No BS podcast because it's not, right? And so you can't also treat this that you govern AI the same way that you govern technology. It's not just about, you know, active directory permissions. It's about everything else that comes with being able to ensure that you can actually move it forward, right? So there, of course, tech components, architecture, definitely, but it's not only technology. And similarly, data. In my opinion, data is a foundational component. We raise the data. We do, right? And so the processes do, with the technology and the tools. So data is really a critical component. And we can talk about what does it mean for AI ready data? I have a lot of thoughts on that. But, you know, data is going to be the foundation. And the more data we have, we think it will be better, maybe, right? But you've got to be intentional about, you know, the job to be done. What are you trying to solve? That'll really, what are you trying to do with your AI? All right. Saying a lot of notes. We're going to unpack all of this. Let's do it. Yeah, you go Tim first. Yeah, I think one thing that's cool, just an insight real quick, is that I like that you, you know, a lot of people talk about like, you know, people process technology and, you know, data or strategy or insert your like favorite kind of fourth item in there, right? And a lot of times people kind of think of like people as the people part, right? But I think what's really cool is that you're thinking about how there's a really strong human element or like a human centered aspect, which I think is bigger than a human element to all three, four, five of these kind of whatever these pillars are. There's a huge human element. So I think that one really really important thing that a lot of folks need to think about And then a second thing is actually a little bit of a question to you So I know some people think that AI governance is a little bit more narrow right And they kind of think of it as like, well, you know, there's emerging AI regulation. So they're thinking about it from like kind of a compliance lens. And then they're thinking about it from sort of a kind of like a safe adoption kind of lens, which is like, we need to teach people about AI, we need to make sure that things go through a committee or we have process, right? It sounds like you're thinking about AI governance in a much bigger way. I was wondering, can you talk a little bit more about that? Totally. So you're spot on, Tim. I think that's a really sharp observation that in my view of AI governance, we're trying to help organizations and individuals, right? How I use AI in my home. I just had one of my kids go through and dictate a story, which they did by just pushing a button to record and it converted it to text. I popped it into one of my AI agents that I use. And then, you know, we created a little book that was seven pages long for the kindergarten class to read, right, about, you know, and it created images. So, right, there's the how I govern it at home, there's the how I govern it in my professional sort of utilization of it. And so it's not just AI regs and safe adoption, in my view. Those fit maybe very nicely in terms of the process bucket, But really, even AI regs and safe adoption is also about the people and how we want to go about using things. So think about anything from students who are going to their first internship. I don't know about you. I've been the mentor of many, many, many very now successful adults through internships. And nowadays, I bet that no matter what the role is, they're going to be using some kind of AI. Do they know and understand that, you know, there are certain things they should be using, certain things they shouldn't be using, and certain things that they should still be checking and using in their brain? I just had another client who really went through and did a little bit of a sales push, but the AI tool they were using conflated some results. And so it was not wrong. It just wasn't quite right. And we caught it after we did the first push. And so now we're thinking about place the right sort of rules. And it's not just about, you know, governing, again, the AI output. It's about actually having and putting into sort of good hygiene best practices around how we work, how we work with each other and how we work with these other tools and the information they give us because they can help us. But we have to also understand where we are still the we, right? And I think that's going to be sort of where we see a lot of organizations if they win quickly, right? Because speed is now picking up. Think about, right, like email. Email was supposed to do what, right? It was supposed to make luck easier. You can now work four days a week. Like I now am on email on multiple devices all the time, right? And anyone can reach me whenever. And it could be the school. It could be the, you know, in-laws. It could be the right. I mean, it's total chaos, right? And so it has made me busier. Has it made it easier? I guess I no longer have to wait for a letter to come in the mail. But I kind of would prefer that sometimes. Right? Give me one day. I'm pretty boring these days. It's like, you know, I want interesting mail. Right. That's right. I'll send you something. All right. Yeah, give me your address. I want to go send you a letter. But this is the thing, right, that, you know, I think the organizations who will really see the movement ahead will figure out a way to complement the humans. First, figure out how to get that invisible stuff out of their heads. Right. And really get it down to do we even need to do this? Are we just doing it because it's how we've done it? And should we really be reimagining it? And so that's really maybe again, maybe the reimagining piece is that AI governance view that I have. because it's not just the same way we've worked before, right? It's not just a new technology where then, okay, checking on how, you know, regulators, where they allow us to send data, right? That was a big thing a few years ago. Now we've got data. Now we've got cloud. Where's the cloud? Where's the base? Where's your data? Can it go into the cloud that's in a different country? Can it cross borders, right? AI, you know, is more of that because it's so encompassing. It can take a lot of different things. And so you have to be intentional with your use cases. You have to know the problem you're solving. Years ago, I tell the story over and over again, years ago, many, many years ago, when data science first started to emerge, I had a colleague who said, hey, Victoria, our team should work together. And I was like, that sounds terrific. You know, what problems, you know, what are we going to work together on? He's like, well, we should do machine learning. And I was like, no, like we do that. But like, what problem are you trying to solve? You know, I think it's been over a decade now. I still haven't heard the answer. I think we've all moved on from those jobs that we had back then. But, you know, you've got to really understand that it's, you know, some of it is great because you have this balance of what do I need to do for business continuity? So safe adoption, AI regulations, right? Business continuity. How do we make sure that we deliver the core of what we need to deliver? Although that too might be reimagined, right? And then there's the sort of how do I spend the right amount of time in exploration innovation, which is where POCs have lived. But the POCs need to somehow get into the business continuity step. And so that's, I think, where we are in many places right now. We're thinking, you know, I'm trying, I'm testing, I'm doing. But if I don't have an intentional plan or a specific, you know, process that I'm going after that I need to reimagine that needs to work with a variety of different tools, then how do I really get there? And I think that's also what we saw with the FDA warning messages that said, you know, this wasn't governed, you know, we don't have enough transparency. And maybe that's because the organization itself, you know, also needs to get up to par. I actually haven't seen that FAA warning thing. Is it like, well, can you share a little bit more about what that, what that, what that was? So it just came out last week. I haven't seen the original documents. So I just saw a report about it. But maybe it's something we can pop in the comments and others can take a look at it. Okay, cool. Because I think that's interesting, you know, because obviously we see that with like, you know, cigarettes and things like that. And, you know, it's this is now an era of where like, you can have, you know, information health challenges. Yeah. Oh, so well said. Maybe that'll be the soapbox next time. So the takeaway I'm having up to now is we treat the term AI governance first of all because it has a term governance we usually treat it more as a regulations right and then people coming from the data governance so that's where it's come from we've been talking so much that governance is really more should be more like an enablement and so forth right that's our analogy is like the brakes in the car is to slow you down but driving fast safely but what i'm really realizing here with this discussion is that it's ai or ai governance is or that term for now as you said now i understand it's an encompassing term that that it really takes everything, which is like, not just all the technical aspects of the regulation aspects, but it's like, oh, wait, there is just, it's how do we, how do we live in this new world where these agents are, they're all over the place. And it's not something that we're going to go do with the fixed approaches that we've always done before, because we are literally in this new world where like now it's the speed for how this is happening, where anybody can bring in like a particular example, like we have all of these agents or systems you can use per the setup in your company, but hey, you got the stuff on the side that you're doing with your, like, how are you going to deal with that? You're not going to go stop with that. Like, what is, we're evolving super, super quickly. Like, this is super fast. And I think AI governance is that just new mindset of how to go deal, how to go work with now these AI agents, companions that we have, and they're they're here already. And we basically are behind in a way of setting this up, figuring this out. On the data side, right, we've talked about data defense and data offense, right? And so the data offense is that like protection piece, right? The data, sorry, the data defense is the protection piece. The data offense, right, is how do we optimize the data, right? And so, and optimize the, you know, value brings. And so I think AI has that kind of similar value risk kind of orientation, if you will. Again, an AI here does include data. AI includes my basic, like, you know, I've got a PhD in biostats, like logistic regressions are now somehow AI. Automation, a lot of people put AI, you know, and automation together. It depends, hopefully not too many. But, you know, this idea of experience compression is kind of happening with AI, or just maybe a general compression. We're able to somehow move faster to your breaks concept. And on the healthcare side, the goal is to move as quickly as possible in a safe and protected way from identifying your target, the one that you want to pursue, all the way through your drug application, your submission to regulators. And so speed has to come, but it can't come at the expense of certain things, right? So those kind of frameworks or guardrails within which you can innovate and explore exist. Those are evolving, right? That's the AI regulations piece, but it's also the standard regulations piece. And so I think there's that kind of component. And I think there's also a lot of boards of these organizations, some of, you know, that we're a part of, that also need to get up to speed with, okay, you know, what is now my role And what kind of questions do I need to be asking in this day and age? And what should I know about, you know, as the models evolve? Because with every new model right there going to be some potential new challenges like you know of where does my organizational vulnerability lie at a corporate level Let say right at a big sort of you know the big picture and then where the individual all humans, right, because it's not unlike compliance than the other day. So there are similarities we can learn from the past. The key, I guess, with AI governance for me is that we're going from a place where we were really good, well, hopefully, successful organizations were good enough at minimum viable product, right? Good enough at the retrospective, the sort of look back, the, you know, what happened before, maybe some pattern identification, right? Of what did I know before? How did it work? Through to now really pushing ahead and pushing forward and saying, hey, I can have more of what I would say prospective, predictive as it ever may be, right? Prescriptive even using some of our analytics ladder terms, right, I can have more of that step forward a little bit quicker, right? It's not, you know, behind the firewall of a data and analytics team only. But yet there may be our individuals who don't have that same data governance background, who don't have the same understanding. I remember just even when stepping into data governance, not, you know, a lifetime ago, but within this lifetime, I had some wonderful colleagues who said, hey, Victoria, like we built this really cool ecosystem. It's a space where you can, it was a catalog, okay, right? It was like, you know, you can find data. And I was like, that's awesome. And they're like, and then you can just download the data. And I was like, whoa, breaks, right? Like what kind of data? What are the rules? Like, you know, because we've got data that people have given their last breath for, right? You've got to be really, we treasure it, but we also have to get the most out of it because they may not be here and that's the legacy that they leave, right? So which is another layer of the human aspect, right, Tim? It's really, really important and it is a big treasure for us. And so it can unlock, but we've got to really evolve to be able to operate appropriately in a dynamic environment. And that's not like we've seen it before. One of the things that you brought up a couple of times already, and we've had this discussion with other people, is processes need to change. And you said something like we have the invisible processes that are in people's heads. How do we get those invisible processes out of people's heads and then figure out what they are and start kind of changing processes? What's your experience or what are your thoughts about this right now? Yeah. So actually, that's an area where I think AI can help. So I'll give you one example of something that I've done. So I have one organization that I'm on the board of, and we're redoing our strategic plan for the next 10 years. And so, you know, I have not been on the board for the past 10 years, and I didn't, was not involved in the previous strategic planning development. Now, again, do we do a 10-year strategic plan and how adaptive is it are all great questions to be covered on another day. But in this instance, right, what we wanted to do was really understand and talk to the CEO and understand sort of what are the opportunities, visions they see in the ecosystem as sort of a foundational starting point for the strategic plan for the next 10 years. And so my thought was, let's create some kind of GPT that will interview them. Right. I can go ahead and I can gather. I'm not, you know, a strategic planning expert. Right. But I do and have read a lot about it. So I know some of the places that I would go to get the good questions. I know I've seen interviews on a variety of different places, right? Different reports published. So I gathered and grabbed a bunch of the resources that I liked. I popped them into a tool. I then used a rag model because I wanted it to be focused, right? So it was, you know, training on those specific resources. It created for me specifically a set of questions, a framework that I can put into. And then this was a, they're at a Google shop. So then I created a Google gem for them, right? That then the CEO could do not just by having to talk to me, recorded, transcribe the notes, right? Which is all AI enabled, right? But this way, you know, and there were still, and I sent it off to, and then the CEO was able to do that and have the conversations. And I gave time because there might go back to certain things, right? Have an idea, have a follow up. And then, you know, I had to do some manual steps as a copier whole chat and send it to me. But, you know, over email. But, you know, so still opportunities out there. But it was a great first step. And so I think this is really that space where I would say for processes to evolve and get invisible things out of people's heads. we have to first understand what is it we're trying to do, right? In the healthcare space, we hope eventually we might not even need, you know, as many people on a standard of care arm or placebo controlled arm, because of course, right, they may be randomized to get no benefit better than what is available today. And a lot of people join trials with the hope that they would get on the randomized to the arm that will have, you know, something new that might be more effective to them than standard of care. But until that day, we still have trials, we still have people, and we need to get those people to places. So there are still going to be some of those in our current and maybe for the, you know, the upcoming time point, jobs to be done that we know about. And then, right, you know, you start, and I think we've seen also, historically, right, we would call it business architecture, but some of these business process type conversations happening in organizations where we're trying to map those things and then to exist. Now, that also, I hope, will get us to a place where it's almost like manufacturing, right? You know, you could manufacture things by hand, right? This person does this, this person, or you could do it more on a manufacturing line. You've got them specialization, right? We didn't know where that was going to go reading back in the history textbooks. I'm not a dinosaur. I wasn't alive then, but you know, that's where I think things really are evolving. So we don't exactly know maybe what that new process will look like, but don't change that process. There's also no reason for people to change. They'll still do their invisible steps. They'll still have their checklist on the side. And as we're seeing also workforce generations evolve, right. And we talk, I remember one panel I talked about, what do we do with the new people who are coming in who maybe don't have those invisible steps in their head. How do you mentor them and train them and give them the knowledge, which not, but may outsource to AI. And how do you get that experience? But then how do you also get that experience in your head? So there's a lot of tools today that can help us do that. But that's also where the human element comes in. It comes from conversations. It comes from people knowing information, and it comes from actually being able to capture it and then break it down into those pieces of what you and I would call information and data. I love it. That's a good insight here. And I think what's interesting to me is I think that, you know, capturing this tacit knowledge is going to be probably a big focus for the next couple of years here for a lot of organizations. And one of the biggest challenges is going to be getting people excited and engaged in that it's not like mining them to try to replace them or automate them or something like that, that actually they're seeing the tangible benefits of the knowledge and the actions that they're doing that are maybe by paper or happening outside and seeing the value of how AI and these other things are going to help them. One thing that you said, as maybe one of our final questions here a little earlier, was you kind of in passing mentioned AI-ready data. And then you're like, I got some thoughts about kind of what that means or if that's valuable or not valuable of a framing. So I'm curious, what's your take on AI-ready data? Is that the right framing? And if it is or isn't, what does it mean? How do we take action around that? Yeah. So one of our common network friends has shared a lot of thoughts on, you know, AI-ready data can be at times sort of BS because, you know, what is AI-ready data? Look at what LLMs were able to do. They just, you know, scrape the Internet and now look at this magic that we have. And as much as I want to believe in purple unicorns, I do know that magic is a little limited to our people energy. And so I don't believe that AI will be some kind of magic bullet, at least not without the complement to it. And so AI writing data, in my opinion, is not so easy. It's not as easy as scraping and grabbing information up there. In language and text, it may work OK. And I think we've seen some strength in our own examples. But when it comes to numbers, what is five in column three? Is it supposed to be 0.5? Is it supposed to be 0,5 or 0.5, depending on which side of a pond you live on? Right. So when it comes to the data that we have, data that's collected from images, data that's collected from patients in clinical trials, data that's captured in our medical records and our claims data that's not the free text. That data is structured, semi-structured at times. That data you know needs context It needs you know we need to understand what it means We need to think about you know how knowledge graphs can help us What is the ontology and how is my ontology the same as your ontology or different than your ontology How is the data going to flow through the ecosystem of my organization? Every organization, for example, deals with finance and numbers. That's a great example. I don't know about you, but I see a lot of finance organizations working in Excel. Terrifying. But again, time for yet another day. But that's a great experience. One of the world's most popular BI tool, right? And, you know, it has its place and I do love it. But, you know, no audit trail. So for our ecosystem, like that doesn't really work, right? No audit trail. So you can't tell when someone changed a number. You don't know if there's a mistake. You know, there's strengths and weaknesses that go with every tool for its job to be done. And for the job I need it to do, that's not the right tool. And so as we look at this data, you know, we have historic data, but the historic data needs to be, you know, things are in different standards. Take weight, right? Weight for one study that had most of the sites in Europe might be in kilograms. And for, you know, another study that was mostly in the U.S. could be in pounds. And so you can't just combine weight, right? Then you need to figure out, well, what is my translation, right? So then, you know, there's all of these things. So I think context is going to be a really critical unlock for that, getting the data AI ready. And then there's just so many topics along with, you know, how is it appropriate to use the data? Where do I put it? What do I do with it? How do I have the layer of, you know, being able to operate quickly with it? Do I point to the data? Do I point to the curated data that now has some, you know, calculated fields in what I want to do? It's a big topic. I don't know that we have enough time to cover it, but this is where I would say there's a big difference when you're talking about data that is numbers, that is in some kind of structured, semi-structured form that in and of itself exists. and without knowing, right, what it means and how to understand it, you can't really interpret it. It doesn't really have the value just by being there. And this is where we've seen in our industry a lot of conversation around data as an asset, right? Because that data in and of itself, of course, is important, but it's data and. Right, right. And I'm becoming, well stated, And I'm becoming increasingly convinced that AI-ready data is really much more about, do you have the sufficient context around the data for the particular use case at hand? It has less to do about the data itself and more around the context that's around it. Yeah, and that's going to really depend on the use cases, on the industry and stuff, because what you're describing, given your background in pharma space, is like you need like that's super crucial. But for other industries, that may not be, right? And then so it really depends on that. Wow. I can't believe like I think we need to have another another podcast session. There's so much like the four things that you went in. I wanted to unpack all those. We didn't even get to that stuff. But we took a lot of notes. Takeaway time. Tim, take us away with takeaways. I'll start by saying that we started off with honest, no BS, AI governance. Go, Victoria. And you kind of started off by saying that you love the term because of the fact that it can be very encompassing and a gateway that can unlock a lot of different conversation about AI, and not just the defensive aspects, but especially the bigger kind of value aspects and really finding what you call the big wins. And Juan mentioned, everybody's doing these POCs and things like that, and how can we drive operational value and production use of AI. And you said that in your experience, especially being in the healthcare industry in space, you've seen a lot of these little pop-ups. And when you think about how you get the most value and avoid what we're going to maybe see more of now, these FDA warnings for healthcare organizations using AI in potentially impact hopefully could be dangerous or risky ways, it's really a focus on humans. It's about human-centered digital transformation where people process technology and data, not just the people aspect, but all four of those aspects have a huge human element. And people, culture, you have to do this very intentionally. Process is very human. Technology, it's not just about the actual AI itself. It's about how you're using it and the bigger picture that it fits into. The human architecture and data, of course, has a very human element to it as well. Jobs to be done, intentionality and context. So lots of human aspects here we need to focus on around AI governance. And you really kind of pushed for a bigger kind of tent for what AI governance can represent, that it's about helping organizations and individuals not just be compliant and have safe adoption, but to really ensure that we have the right impact. You know, business continuity being reimagined, business value delivery and innovation being reimagined, you know, not just sort of what we messed up in the past. Hey, the data science era, like machine learning, let's figure out what to use it for, right? Problem searching or a solution searching for a problem. You know, let's take this with a very solution forward kind of mentality. So much more. It was just the tip of the iceberg of my notes. But Juan, what were your big takeaways? Well, I like how we discussed a little bit about the kind of the passing we've done with data. Like you see data about being defensive, offensive, right? Defensive compliance and stuff, offensive optimized. We're right now in this whole era of AI that is, this is applicable too, right? We need to move quickly as possible in a safe way. Speed can't come at the expense of other things. But what changes a little bit now is that we have like, think about like vulnerabilities around this organization. and even as people, as humans, like that is now something that we were not doing before, like in the whole data space. That's what really changes here. So we were really good at the retrospective, understand what worked and what didn't work. But now we really need to kind of figure out a way to kind of predict what the future is going to look like and not the way we've done things in the past year is going to help around this. Which is interesting. We've talked about like, how do we get these invisible, the invisible kind of processes out of people's heads? And what you described is something like, I've seen other people trying to do this too. It's like, hey, let's go create the agent. It's going to interview the people, interview the expert executives about what they're planning to go do. And then in that process, you try to extract this thing, extract how they're doing things and the processes around that. The thing is that, you acknowledge that we don't know what that next process is going to look like. And I think that's kind of what's really interesting. At least we need to understand what we have, how we're doing things today, such that we can go figure out how we can improve these things. And then there's those aspects of like, people are being onboarded. And like, how do they know those invisible processes? Like, this is actually how to get help. By the way, It's not just people, even the agents themselves are onboarding agents. They can do that. And then we wrapped up with the whole AI-ready data. And I think it's a very clear separation of like when you're doing things around language and text. Yeah, maybe that scraping type of stuff is going to work. But numbers, right? When you see five in a column, what does that number mean, right? That data is captured. And I love what you said. This data is going through a flow through an entire ecosystem of your organization. Like that's something we need to understand. It ties back to like how all that flow of the data ties back to the business processes you need to understand. You need audit trails for things, right? For example, data have different units. You need to understand what those units are and how to translate between them. How is it appropriate to use? How can I operate quickly? This was a really interesting thing. One is if I want to operate quickly, can I go to the raw or should I still go to the curated? And there's a change there. There's all these little different things. And all of this really defines what AI-ready data is. And it's not just a, oh, have a high data quality, whatever. It's like, this expands so much. How did we do? Terrific. I think what I'm hearing sort of from both of you is that in a way, and again, right, I studied the medical aspect of our world. Think of it maybe like a muscle is sort of the light bulb I just had as we were talking, right? It's a muscle. And you have to treat your muscles well, right? And you have to, you can pull a muscle, you can hurt a muscle, you can strain a muscle, muscle, you can rip a muscle, right? But if you, you know, treat your muscle well, if you feed it with good things, um, and an occasional cocktail, right, your muscle will, um, you know, be able to push you harder, make you run faster, you know, um, and help you, um, you know, achieve, uh, a lot of great ambitions that you have. So in a way, I think that's my hope for this space. To wrap up quickly, what's your advice, who should we invite next and what resources do you follow? So I follow a lot of folks on LinkedIn. I follow you guys. I love the podcast. I also really like the Data Chief. I think there's a lot of high-level topics there with Cindy, and there's just too many to name. Next guest, goodness. Geez, maybe some folks who are really in the space of, or someone who has, you know, used AI to actually extract that knowledge and has captured an invisible process. But I have a feeling that we're all still on that journey. And advice. I think, you know, Tim, at the start of this, you know, you had mentioned the word curiosity and learning. And I think to me, that's really where we are. We're learning. Things are a remix. And so that is the opportunity for us to sort of stay curious and stay open. And I think, you know, we have to be cautious. But in a way, right, it's no longer if we use AI, right, in our world. It's really how we use it best. That's it. Victoria, such a pleasure. I'm looking forward to doing a part two of this because there's so much more to unpack. Thank you so much for your time. Cheers. Cheers. Thank you.