AI変革に必要なのは技術だけでなく人とプロセスの再設計

Welcome to the Pulse of AI. Insight Kitchen was built around the singular focus of turning the promise of AI into operational reality. He has a PhD in applied math and immediately before founding Insight Kitchen, he was the head of Accenture's West Gen AI practice. I hope you enjoy the conversation. Let's get to it. Well, hey, welcome to the show. Hey, Jason. Great to be here. Thanks for having me.

Of course, of course, I'm excited. Before we dig into, you know, today we're talking about sort of AI transformation in the enterprise, right? You know, really, you know, really focused on, you know, mid-market companies and, you know, even more specifically, if we get to it and dig down into the supply chain side of it, right? Which I know you have a lot of experience in. But before we get to that, personalize yourself to my audience. Tell us a little bit about yourself, your background, how you got there, what drives you? Sure. Again, thanks for having me. My name is Russell Halper. I founded a company named Inside Kitchen, and we focus exclusively on how do we help companies integrate AI into their supply chain. I go back far enough in my career. I actually did a PhD in applied math. I was studying chaos theory with somebody who was actually famous for creating the phrase chaos theory, which somehow feels appropriate to supply chain sometimes. But it moonlighted at my university taking a class in the business school studying operations research, or what probably more recently be called supply chain data science.

Wow. And I enjoyed that. I had a chance to work with one of the large parcel delivery companies and how they do some of the vehicle routing and some other projects, and decided that was the direction I wanted to go. I left from there and joined one of the large consumer goods companies in the U.S., working inside the supply chain, doing these types of things as a hands-on practitioner. And there I had the experience of both working from doing, like, the math to actually going to the factory floor and spending a lot of time, you know, on production scheduling on a variety of products. Left there, joined a consulting firm in Palo Alto. You know, back in the day, if you were in data science and organization, there wasn't a lot of places to go, but if you were in services, you had a chance to be able to work with all sorts of different companies. We had that opportunity to work across semiconductor, automotive, retail, consumer goods, and a variety of others doing all sorts of problems. That little firm was acquired by one of the very large consulting companies.

And so I had a chance to be there and be there at the start of Gen.AI. They wanted to build that practice. So I helped build their West Coast generative AI practice. But ultimately, I felt there was more opportunity to help companies the best way we could by being able to start a small, more nimble group that we can really focus on. How do we bring, get value out of this latest technology with AI? And that's what we do. So that's a bit of my arc. Thanks. I love it. I love it. It's interesting because, you know, I mean, you just looked at the profession of your career, right? And what we call AI and what has been AI traditionally encompasses a lot of different sort of approaches. Right. Deep learning, machine learning, you know, just, you know, computer vision. There's just there's just a lot that sort of falls under under that category. But and if you went and talked to a lot of CEOs, you know, and I interviewed a bunch, you know, pre-ChatGPT. Right. My podcast predates that. And if you ask them, hey, are you doing AI in your company?

A lot of them are like, I don't think so. No, we're not. We're not doing this. So then ChatGPT came out. It sort of overtook this, you know, and became AI in people's minds. And and so now everybody wants to do AI. Right. And, you know, pre pre ChatGPT and you and I have talked about this. The previous forms of AI. Right. Were applied heavily to the supply chain. Right. It was one of the major application areas, you know, really trying to, you know, optimize and predict. I mean, just everything, right? Right, yeah. But supply chain is slow to adopt gen AI. And let's start there, if you will. Frame that. Why is that? And what do you think is holding it back? And what are some of the opportunities there? So I'm going to give a validation point to what you're saying.

We just had a big conference. It was an 11,000-person AI conference. They had 1,000 speakers, three talks on the supply chain. And so if you think about that as a natural sampling as to, like, where you're getting penetration with generative AI, generative capabilities, it feels behind. And we often talk about this concept of operational AI, which is how do you tie AI to the physical operations that you're doing in an organization. So if you're using at one end of the spectrum, you know, Claude or ChatGBT to write an email. There's a lot of different ways you can write an email. At another end of the spectrum, and this is less agentic, but if I'm using – if I'm doing something in the world of physical AI, I'm operating a humanoid robot or grabbing a car, I'm getting real-time feedback from, like, the physical world around me. We're talking about supply chain and operations. You have this disconnect where I have real physical constraints of things that are actually happening on the ground in an organization where, you know, I can't hallucinate inventory is somewhere that isn't.

I can't send a truck to the wrong location, but I need to figure out how do I use AI to drive value in those operations. And that's where it gets hard because you have highly constrained and customized operations to organizations that may be different organizations selling competing product on the shelf, yet operate very differently. And those organizations have to figure out how do they use AI to drive these, again, highly constrained, complex decisions that involve direct cost. And so it's a longer pull and a tenth to think about how do you adopt some of these technologies than say if I'm a developer writing code. Right. And that's, I think, some of the constraints that we're seeing organizations, as much as they want to use it, are struggling to figure out how to adopt it. But it's interesting. I wonder if there's going to be a change. Supposedly, since 2022, it's the year of the agent, right, at GenTKI. And, you know, it's obviously taken longer to adopt for everybody around there. But we're getting closer and closer, right?

I mean, do you think that agentic AI will change that in the supply chain? Because, I mean, just look at the supply chain, right? And you would think that agentic AI was almost custom made for the supply chain, right? And so do you think that's going to change, right? I understand a lot of the ways that they've been using gen AI up until this point. But from a workflow agentic perspective, what are your thoughts on that? So when you're talking about your logical supply chain, planning, complex decisions that involve quantitative tradeoffs, it's hard to make, at least with today's capabilities, the agents, they don't necessarily work as well as some of the old capabilities. But what we're seeing more often is the marrying of those two things, right? You were talking before about like machine learning, right? Neural nets, things that have been used for doing numerical prediction, which are using in forecasting all the time. And that's a well, like a thorough battle-tested, valuable thing in supply chains.

What we're seeing more are organizations today marrying those with agents as analysts, as control layers on how they interact with different scenarios, and those two things working together. And so I think we're going to continue to see a penetrate, like more and more penetration at that logical level as these things weave together. On a more execution level, I think there's a lot of opportunity around how do you think about automating operations, understanding where should a person get involved? What are the things that you can now hand over to an agent? And like that's another ripe area where we're seeing we're starting to see more and more penetration. Yeah. One of the other things about supply chain is you have all these islands of systems often in the supply chain that don't all talk to each other. You have your ERP and your TMS and your WMS and your planning systems, and the planning system might be connected. It might be disconnected. And oftentimes when you want to ask a question, like, why did I miss an order?

You now need to look across all these different types of systems to be able to get to those types of questions as well So again it an opportunity for a general of AI for agents to be able to help answer those questions much more efficiently than we can today. I think it's also a challenge for organizations to figure out how do they get all the data together in the right context on top of all of it to be able to leverage for some of those things. That's interesting because, you know, it leads to my next question, right? And you and I have talked about this in the past. And you really, in order to really benefit from AI, right, especially Gen AI, right, especially agentic AI, right, you know, workflows and this, you really have to transform your company, right? You have to transform, you know, organizationally, you know, from a leadership level all the way through. So can you talk a little bit about that, the challenges with that, the necessity for that, right, and the challenges for that,

and follow that on with this concept of if you're going to change, right, and if you're going to transform, it's not just about augmenting the task you're doing today. It's about rethinking processes, right? It's about rethinking how you do things, kind of like what happened when we started with machine learning and deep learning years ago on predictive analytics and data science, right? A lot of supply change radically changed. I mean, sort of unpack that, if you will. So an example that I'd like to talk about is with agent capabilities, a lot of the conversations, especially early on, where can we replace people or reduce our headcount by leveraging AI? And whether or not that's a good thing to do or a bad thing to do in certain capabilities and organization is a whole interesting question. But let's just assume for the moment that circle, and we're going to take planning as an example. If I think about a large multinational apparel company and how I'm planning demand in that organization,

I like to pick demand often, you'll hear me go to it, because I think it's just a relatable concept. But this is true for inventory, allocation, merchandising, et cetera. But let's just pick demand. So oftentimes when you look at the way the demand plan is divided across the organization, I might have it broken up by APAC, EMEA. And then within North America, I might have my different franchises. And then within those different franchises, I might have, say, men's, women's, children, accessories. And then in, like, men's, I have tops and bottoms. And then I might even start to delineate more than that. And I have a demand planner there who is, you know, North America, men's, you know, et cetera, et cetera, et cetera. And their part of the world they're responsible for is actually delineated and sometimes more by that part of the organization than, say, the actual capacity that they have to plan. And they end up being full because there's all sorts of things that happen.

But if we wanted to come back to the premise here of we want to automate work and reduce our headcount, you have a structural issue that even if you can make the work take half the amount of man hours to get done, you still have to think about structurally how is that related, how are you thinking about the divvying of planning, and not just with that person, but all the people they're then interacting with in the organization. So if they're demand planning and they're interacting with supply and inventory and merchandising, and that's all structured that way, there's a certain number of touch points in the day that those people have, and those touch points still take time. So there's structural components that come into play when we think about that example of how to do it. And it's just like, to your point, even with data science, the way we might fundamentally think about approaching processes can change. And it's an interesting opportunity, but it's a challenge that makes it not just a technology challenge, but also a process challenge. Yeah, it's interesting. So if I was to just take you, right, and drop you in to, let's say, a mid-market company with $500 million in revenue, 300 employees, whatever, you know, something in that range, right?

and I dropped you in, into the C-suite, where would you want to start? How would you work with the CEO and say, this is the first thing we're going to do, and then this is how it's going to progress? Well, one of the things I think about mid-market companies, which is so interesting right now, is the opportunity relative to some of the larger guys in terms of being able to move fast with the technology. within our client base, we just see the rate of adoption in those $200 million to $1 billion companies, dramatically outpacing what we're seeing in companies in the $5 billion to $40 billion range. So when we think about where to start, I still come back to what are the big areas of addressable strength? What are the actual problems in your business that are things that are either barriers to growth or barriers to actually driving profitability? And how do we think about tackling those?

And I think of it as a business question first. If I'm running an organization and strategically I'm trying to think about what are the three things I need to do to make my company better? All right, let's start with those three things. And then let's work on can AI help support some of those things? Now, there's the kind of ground-up approach as well. Okay, let's get everybody with their own agents or ability to build agents or give them clout or something to make their work more efficient. And that has value. It's sometimes harder to measure that value. But from an executive level, from a P&L level, there's no reason we shouldn't start with the same place as we would start with any type of technology transformation. What are the pain points in the organization strategically? what do we need to do to get better? And then how do we think about if this new technology can help that? That's a great thing. A guy I know who does a lot of work with this, like you, his thing is CEOs and similar questions and says,

where do you want the business to go and what parts of your business are actually holding you back, right? And then going through and saying, okay, and learning sort of what I see as one of the challenges is that, you know, sort of pre-chat GPT, you go in and talk to a CEO and then say, oh, AI is all this great stuff you can do. And they're like, I don't believe it. That's science fiction, whatever. And then after chat GPT, every CEO is like, I think this can do absolutely everything in my organization. It's almost talking them off the hype to say, well, it can actually do that, but it can do a lot of really amazing things. So how do you have that conversation when you sit down? How do you do that? I'm also wondering if you had the conversation where they're like, my organization is behind because we haven't adopted the level of AI that I'm seeing in the world yes, yes and it's some industries, AI has got tremendous traction others, it's got more of like it's Instagram moment where you're seeing everybody's highlight reel and more organizations

are at the starting block than you might think or supposedly they're using AI because they have a license for ChatGPT and their employees are using it, right? Yeah. That's using AI, but that's different. Or even if they don't have a license for ChatGPT, all their employees are logging in anyways in the ChatGPT and taking all the confidential information is thrown on the web for them to use. Right. It's the new personal device problem, right, with a lot of these companies. But tenfold, magnified. Yeah. But how do you see that, though? Like, how do you have that conversation where you really have a base set conversation with the CEO around, okay, here's actually the areas where you can deploy, right? They say, these are things that are holding me back, et cetera, et cetera. And you say, okay, well, let's talk about that. How do you set that? Does it require somebody like you coming in who's been doing it for 20 years and has a Ph.D.?

Or is it dumbing it down for the CEOs? Or what does that conversation look like for you typically? And what holds them back, right, at a C-suite level? So I feel like there's a lot of flavors in this conversation. And that's where we start dissecting the world. So, you know, there's the organizations that want to be the early adopters, the ones that want to be the fast followers, the ones that are a little bit more skeptical on how they invest. And at that C-suite level, I think all those things need to be addressed a little bit differently. But when we hear from people that think about, hey, I can replace everything with AI, part of it also is how much risk you want to take on the organization and what's your capacity to invest. because even if you can replace large amounts of your workforce with AI, what are you losing in terms of the tribal knowledge in your organization, right? Your capacity to learn about things that people might learn more easily about but AI is going to have a harder time picking up You know what are the conversations people have in the hallways How do you make sure you still giving the human touch to your customers at least where you want to

I mean, there's some places where automating that is important. There's other places where you have very personal relationships, and that's actually what's driving you as an organization. And so part of it is realizing also where does it add value and also what are the things you might lose? and then to your point, what does it actually do well and what does it not actually do well? And that's, I think, a harder question for folks that are not in the field to always conceptualize around, like, throwing it at the wrong thing. It can, like, really confidently give you wrong answers, which is in itself another risk organizations can take on instead of looking at how do they bring in this new technology. you know i guess there's like you said something a little bit ago that said it's like having the conversation and i want to push back on this a second and you said the past conversation they've had about past technology transformations right and one of the things you know one of the sort of

the mantras that i have is when i talk to people is that it's actually fundamentally different than past transformations, right? Because implementing a CRM system, that's something, right? And that's interesting, but it doesn't fundamentally bring into question the core processes within your business, the reason for those processes, if there's a need for those processes anymore. It doesn't have you think through the workflows within your organization, right, and the way it works, the workflows both internally and how you interact with customers. And I guess just the potential is different, right? And so I think one of the mistakes is, just sort of round out that thought, is that a lot of CEOs, because as we know, a lot of AI initiatives aren't bearing fruit for a lot of CEOs.

And there's a lot of spend out there, right? And they're seeing pilots that fail. They're not realizing ROI. But I think part of that back on this organizational leadership thing is that they entered into it thinking of it more as a new IT initiative as opposed to a this is a real opportunity and a demand for us to completely rethink everything within our business and where to apply that. I'd love your thoughts on that. So I'm going to agree with you. The potential is tremendously different, right? It's something that's a whole different animal than the things that came before. I was trying to draw similarities. One of the things I think about is there's fundamentals like good project management, good ways you're going to run an initiative in a business, you know, fundamentally evaluating what's the business case for doing something. And I think those are still true. And those are the things I think people lose sight of when thinking about bringing this into an organization.

But to your point, because of the new potential of this, right, reworking processes is not just an opportunity. It's almost mandatory if you want to get the value out of this when you're thinking about operations in an organization. I saw a slide the other day from an organization, and, you know, it went from here's our SNOP process today, and it had the classic SNOP process. Here's a nice circle. Even though in reality, if you ever actually map one of these things out, it looks nothing like a circle. But academically, here's the so many steps we have, and we do this cycle every week, every month, whatever that organization needs. And then we want to go to this, and it was a bunch of arrows that pointed into a cloud, and it just said AI. And, you know, organizations have this idea sometimes where it's like, hey, I can just get rid of all my process and run everything in parallel and dynamic and just throw AI at it. And it's, to me, almost the opposite. You need to have well-defined structure to be able to bring in some of those technical capabilities that AI brings.

And it's a different type of thing for sure than like a CRM or ERP or a planning system because the power of what it brings is so different than anything we had before and how it approximates the way that people typically can think about a lot of these problems. But the thing that I think is important that organizations sometimes lose sight of is your business fundamentals are still the business fundamentals. And, you know, if you're selling cleaning products on the store shelf, how do you know that implementing this is going to drive more value, right, in that type of organization? and being really thoughtful about how do we get to that in a rigorous, meaningful way that, you know, even if you're the CEO of this, how am I reporting to the board or our investors that I'm actually being a responsible steward of this technology or creating actual measurable value out of it?

Well, I think, and I agree, I completely agree with that. And I think it also gets back to your point that you've been emphasized a couple times during this conversation on, and I would agree with you on this, is that it's not an employee replacement tool. And too often companies are looking at it like that. That's how they price it in their mind. They're like, well, I'm spending X on labor, so I can spend this on Claude. And, you know, if it plays 10% of my labor, you know, whatever. And I think fundamentally that's the wrong way to look at AI. AI should help you grow, right? It shouldn't help you shrink. It's not a cost savings method per se, right? You know, for instance, I know a startup, and there's a couple of them, but one I really enjoy, that's helping sales. You know, it's really always staged, but it's helping, you know, sales teams. It pulls in all the data for all the reporting that salespeople have to go through in these CRMs, right? So CRMs oftentimes add in more work per employee to a CRM.

And pretty soon, that's an extra 10 hours a week that they're putting on. AI does just the opposite of that, right, where it pulls in all this information. You don't spend time reporting. It already reports, right? It already, you know, does forecasting. It does all of the account information, and it brings it in, right? So you don't sit there filling out these forms hour after hour after hour, which salespeople shouldn't be doing. That's not their skill set, right? Their skill set is go out and talk to people. So CRMs reduce the time that salespeople got to spend with customers, in theory, if you had the same amount of hours per week. AI helps increase the time that you get to spend with customers. And that's a critical distinction. And if you look at that at every function within the organization, that's where it becomes, I think, really exciting. Well, another great example is just thinking about there's oftentimes work in an organization that you just don't have the workforce to be able to get to. If you're a food distributor, typically you have a large number of low margin transactions and you just can't hire a workforce oftentimes to look at every single one of those individually and ensure you're compliant with all the terms in your contracts.

There's not other issues. And so revenue leakage tends to be relatively high in those types of organizations. And so, you know, think about now I could bring AI in to help me identify, you know, where those issues might be occurring. And even if the AI is 50 percent accurate, right, that's all value capture. And in practice, it could be significantly better than that. I'm just taking a fairly low bar as a starting point. but it's another place where, again, what's the stuff you can't get to with a team you have, especially in the 300 person, $500 million organization. My capacity is limited. So if there are places where there's code leakage in the organization or value loss in your organization that I just can't get to because a lot of low value individual transactions, it becomes an incredible tool to be able to go and chase some of those. And it doesn't actually have to be as good as a person then, right? It can get to the point where even if it's part of the way there, it's adding value.

And who knows? It might be way better than an individual. I'm just talking hypothetically as we think about ways that it can actually really impact organizations. Yeah, I agree. In Silicon Valley or the technology industry and AI specifically, you know, we keep having this dream of a one-person billion-dollar company, right? And there's all these bets. You know, when's it going to happen? When's the first one-person billion-dollar company? and that may or may not ever happen, but I think it will. But, you know, there's this, it brings into play this concept of, you know, if you were to be able to start from scratch in any company today, right, in any industry, so I'm going to do insurance, financial services, whatever, right, supply chain you know how would you build it from the ground up with AI I mean you still have a lot of people and you put people in the loop right And humans in the loop But the question is what roles would they have then within that business, right? What would that look like? You know, as opposed to looking at it as purely a task replacement tool or a person replacement

tool in an organization, it should be looked at as how do you 10x your company, right? How do you take your company from today and 10x it by augmenting your workforce, training your workers to do something different, you know, you know, penetrating markets that you couldn't get to before, you know, is sort of like the example of you just didn't have the manpower or is slightly adjacent to what you're what you're doing. Right. And and you look at it and mid market companies, If you take that sort of, you know, idea of, you know, smaller, a little should be more nimble, they should be able to compete. And this has always been the hope of AI, right? That they should be able to compete with the large companies because they can, in theory, bring on millions of agents, which are kind of like employees and scale, right? And how many of the companies that you're talking to out there are thinking that way? And are they even thinking to think that way?

fewer than should be are thinking that way. And the ones that are thinking that way, it's interesting because I find there's a strong correlation between the ones that think that way and the ones that, frankly, are already doing pretty well. And I think part of it is that mindset of innovation and growth and, frankly, just being tactical with good operators of an organization. And this is where, you know, the large companies, we tend to see lots of challenges that AI initiatives oftentimes get bogged down in bureaucracy, budgets, spend. You have IT, you have the business. They're not working that closely together. You know, there's oftentimes competition on how to own it. And then we look across all the different functions that might want to do something versus some centralized part of IT that might own the technology footprint that's underlying the capability. there can be lots of dynamics that come into play that can be unhealthy in terms of the velocity that the technology can unlock.

Now, you could argue, okay, security is important and all these reasons why, and not trying to disagree with that. But just like in contrast to a mid-market organization being able to rethink how we're operating is just incredible. And one of the things I've seen that's the most impactful for a lot of these small companies is if they're able to get their data into a place where they have the right context on top of it, the right harness on top of it, and being able to leverage it and democratize it across the organization. That's been so transformational for some of the companies that we've talked to in terms of their velocity of not just operating the business, but then being able to do the next level of transformation with AI in their organization. Yeah, it's fascinating to me. And you just touched on it. And we're hitting on that right now. Is that at mid market, whatever company, if you're a large company, it's just much harder to do. But mid market companies, in order to really benefit from AI, right, you have to create a culture of innovation, iteration, right?

You know, an ability to take risks, right? And not be punished for that. much. You know, a lot of cultures, you know, obviously, you know, corporate cultures punish you for taking risks, but create this flywheel of, you know, innovation that you need going forward. Because I think one of the core messages that I try to get across, and I know other people do too, to business leaders is that unlike CRM or unlike, you know, cloud, SaaS stuff, this is changing rapidly. So, you know, you can sit there today and say, okay, how do I apply AI to my customer service? And how do I apply it? You know, and look at it as technology of that day, right? But you're always going to be two steps away from, I mean, two weeks away from the next advancement of AI that's going to make a lot of that kind of irrelevant, right? And so then you're always playing catch up as you go through, as opposed to rethinking your business, right? creating this concept of we've got to constantly innovate.

We've got to get people who are forward-thinking, always moving on this, because there's no, with AI right now, I just think there's no steady state in business going forward. I also think you have to learn. And to your point about failure, to me it's not about success or failure. It's about how do we learn and continue to learn and evolve and build that organizational muscle over time. And then hopefully things are set up for success. Like that would be the ideal. But if we think about the pace at which AI is evolving, everybody has a hard time keeping up. It's a full-time job to keep up with what's happening in the marketplace. But that doesn't mean we can't capture value along the way. And so this is a fundamental question of do you want to capture value now or do you want to wait until later? and if you wait till later, you now don't have the organizational muscle that you've built to then work on how do we adopt that. And so it's just this world to the moment where I think there's just,

if we're making this analogy between a slow, large company and a fast mid-market company, it's a similar type of argument here of those that get started and learn are going to figure out more and more how do they integrate into their business, When the next capability comes out, they're going to understand more about what that means fundamentally for them, their operations, the processes, the bottom line. And it's going to be this world where I think there's going to be an increasingly large gap between the early adopters and the fast followers. There's no doubt about that. And plus, I think one of the exciting things about this and one of the bigger differences, one of the sort of stark differences between, you know, AI and let's say CRM systems, right, or some of the latest stuff, is that I do not know a CEO right now that AI is not absolutely central to his or her daily thinking, right? Like, that's what they're trying to figure out.

And I've been in technology for a long time as you have. CEOs were never thinking day after day, oh, the CRM stuff is going to be fantastic. I was always like they get a report once in a while. How's the CRM system going? That's fantastic. But this is AI is fundamental. So it's core to the business. And so it's exciting because I think, like you're saying, you know, that certain businesses can be like employees, going to be like individuals, right? The people who the old IBM saying people who AI is not going to replace people, people using AI are going to replace people, right? And the same thing is going to be true with companies. If you don't figure out how to do this as a mid-market company in whatever market niche you're in, the companies that do are just going to eat your lunch. It's as simple as that, right? And it's not even going to be immediate, right? When you start talking about small differences in margin that can grow over time and what does that mean in terms of their ability to compete in the marketplace, right? It can be dramatic or it could be death by a thousand cuts.

it's just, it's this potential differentiator that can add a tremendous amount of value or it can end up behind. Exactly. It's a tough thing. Well, listen, we're running up against time here. I want to have you back on. You know, I'm excited to have you back on. I mean, we can talk for two hours and there's a bunch of other questions I have here. Before we leave today, what are some, like, last thoughts you would have for people who are, you know, either in process on this journey and not having as much success in their company as they thought they would with AI, or just sort of beginning the process and trying to figure out how to apply it? What are some last thoughts they could take away from this? So I think it's always valuable. First of all, I would encourage them if they're struggling with it to keep working because there's an opportunity to learn, right? It's about building that muscle, understanding more about how it's going to apply to an organization. And sometimes some of the most creative ways to think about where it's going to be don't come from the top. They come from somebody who's in the trenches and they're doing something and they see an opportunity.

So I would encourage them to not just be in a bubble and think, but reach out, talk to different parts of the organization, talk to advisors, try to understand where you might be able to find ways, you know, share what went well, what didn't go well. Try to get that feedback because it is an opportunity that we're seeing in the marketplace, and a lot of our customers are finding ways to really grow from it. But there certainly are lots of starts and stops, and there's lots of smoke and mirrors out there right now, and it is a tough thing to sort through. But keep working at it because I think there is just a tremendous opportunity for those that are going to be in the lead in the space. Russell, thanks so much for coming on. It's been awesome. Thank you. Thanks for having me. Appreciate being here.