Defending Your Valuable Data Assets, with Kristopher Moniz

started, I wanted to share, you know, over the last, we've talked a lot about AI this past year, and what's happening with AI. And in since one of our last episodes, I thought I would share, they've been publishing a lot of the stats around Copilot as it's come out. And so, you know, if you're not familiar with it, Copilot, they let some of the larger organizations in the world beta test co pilot, we may talk about this a little bit

today. But some of the stats that they've published are that 78% of Copilot beta testers felt they were more productive. That's a stat from Microsoft. I read another stat on a DevOps website that 20% of folks using GitHub, GitHub Copilot, which is for developers feel that they are more efficient with GitHub Copilot. And so it's interesting, you know, we talk a lot about efficiency. Why? What's happening in IT? What are the trends? AI is a big one, data efficiency. So the stats

are now finally coming out. We just got access to the Copilot, kind of the copilot full test and we're seeing a lot of productivity increases around it. So more to come, I think you're gonna see Copilot be a bigger impact in 2024. And so keep following it when you do get access to it, play around with it. It's very cool. It's got a lot of awesome features that you can use in your organization. And I think it's going to be a big game changer. So with that, welcome to the

show. We have Kris Moniz from Centric Consulting, he's a data

and analytics practice lead. I'm gonna let Kris introduce himself. So Kris, welcome to the show. Thanks for joining us. It's the last episode we're gonna film before the holidays, so thank you for being our final guest of 2023.

technical services provider. We do all kinds of projects for clients, everything from operational, operational excellence, being able to do process transformation, and that sort of thing through large scale data projects, which is what we do, and absolutely everything in between, custom application development, organizational change management, we generally try to be just really solid partners for our for our clients.

strategy? And where do you start?

technology? And that ends up becoming the beginning and end of it, you will look at those strategies and you see a lot of transition diagrams of current state architecture and then I have the five transition states and I finally see the end state. And that is 90% of their strategy. And all of those things are necessary for really functional data strategy but it is certainly not where you start. It is one of the many things you have to do to get a

functional data strategy. We always start with, what is it that the business is trying to accomplish over the next three to five years? Ideally, when we're dealing with a client, if they already have a three to five year business plan, that's where you start. And very quickly, you turn into interviews and conversations with business owners trying to understand not what technology do they want, they frequently don't care. You're trying to figure out what capabilities are

they trying to enable? What's the business architecture that they're going for in the future? And what's driving need of each of those capability requirements? And then how do you translate that into an overall solution and an architecture that's going to help them get there? And that is where you start wrapping your

strategy. And once you know those things, it becomes a broad conversation about what comes first, what comes second, why, what lift does this capability bring you over what period of time, and how quickly can we make that capability happen? Data strategies have to give incremental benefit as they get executed. Otherwise, they're really pretty presentations that you spend a lot of money on, and they don't get well.

there. In your opinion, how does an organization end up with that being the like, what they're what they're executing? Is there a, is there a theme? Is there a, Hey, this is how they typically got there? Like, how does an organization end there? And what do you see, most likely?

implementations. One of the biggest failure points we often see is that sort of Field of Dreams scenario, if you build it, they will come. And when you take that approach, what ends up happening is you build it, and yeah, they don't come, because it's not built with them in mind. So what we end up seeing is, there's that scenario, that happens all too often. And it usually ends up being highly

technical. The other scenario is, we see an inordinate number of clients will go out and they will look for a very, very, very detailed data strategy that they will spend a lot of money on. And they will make sure that they buy it from a very large name provider, right. There's that old saying of you know, pick your top tier provider, nobody ever got fired for hiring

them, type of thing. And what they frequently end up getting is they get this very, very high level, very, sort of advanced, and well, frankly, very high minded, deep strategy that goes about 1000 feet in, and stops. So it's a ton of very aggressive ideas and this is where you need to be. And there's no action in

it. It stops before it gets to the here's the roadmap, here's how you're going to have to deliver this thing in increments, and the type of folks that you're going to need, the cost estimates and all those things. And sure enough, many, many times we end up getting phone calls of well, we just did this was such and such, and here's what we got, and now we don't know what to do with it. So it's those are the two that

we see most often. And it's unfortunate because they both cost time, money and a basically lost opportunity. And those are the things you could have avoided.

teams. The folks who are going to benefit the most from having access to data and being able to be more nimble about how they execute data and the business when they get it. And find some really key use cases from those folks, that can bring a lot of lift to the company. And then use those use cases and work with your team, just try to think through a couple of solutions to solve those use

cases. And go to your finance team for, hey, we want to tackle this use case, we see a lot of lift here that we can get for the organization, we've got sponsorship from the sales team. And we want to show you a way that we can start to do these things, and give them a proof

point. So when finance comes back in four months, and you've actually delivered something for sales, when you deliver that thing for sales, you've got to deliver not just the capability they asked for, but you've got to deliver it with a definition from them of here's why that capability was valuable to us and here's why that capability is valuable to the company. So when you go back to finance, you can say we delivered it, they're

using it. And then in a couple of months, you can show and here's the lift they said the company would get and it's happening. So you can then talk through look, this is a small example and it's very targeted and it's very tactical. And if we want to start seeing much more strategic outcomes in the longer term, where we're not doing this, in you know, really small increments, and, frankly, in a bit of a vacuum, right, we

just did this for sales. If this was strategic, we would be roping in every part of the organization. And we probably would have uncovered some parallels in a couple of different pockets of our value chain that could have gotten use out of this as well. This is just a hint of the types of things we can accomplish if we get strategic. And then from there, that's where you usually start reaching out and saying, Look, we can only do so much

with what we have. And while we think we may have a team that can maintain something like this into the future, we don't have a team right now that can build

all of this. And that's where you start saying, Yes, we're going to bring in some folks that can help who have experience doing this, who can help turn this your vision into our vision, and then use that to help get you overall a bit of a high level estimate to go back to that CFO and say, This is what we should it's going to take for us to get to a real strategy. And then out of that, here's what we can expect in terms of mid to long term deliverables that will help us get to a broader vision.

the business plan? What are you looking for to say, here's why we really need to have this data strategy. Here's how we build it. Like why does that make it successful?

it's about. Probably the best way I've ever seen it worded and it's a relatively new concept that is starting to take hold in the last 10 years or so is you always had an overall technology architecture and enterprise architecture. And within that you have the data architecture and application architecture.

But about 10 years or so the concept of a business architecture really started taking off and I really love that concept because what that brings to the board is this idea of, frankly, if I own a business, I don't really care what my data architecture is, or my enterprise architecture or my application architecture. I care about what I make, what I sell, and my customers. And if I'm a good company, I care about the team that helps me do all that,

my employees, right? None of those things are my technology. My technology is an enabler of all of those things. So if you don't start a strategy, with your business architecture, which really or your business plan, which outlines your business architecture, which really identifies what is it I'm trying to sell, who am I trying to sell it to, how do I need to support my team in doing so, and how do I measure my overall success, and deal with my

customers? If my plan doesn't start there, then how do I know whether or not my technology that I'm spending money on and I'm investing time to integrate and configure is actually driving any value or even potentially driving negative value in my organization? Right?

How many times have you heard about an application or data solution, anything, any piece of technology in the business where the the general theme of those who were forced to use it is this detracts from my day, it doesn't help, it doesn't bring me lift, it actually makes it harder for me to do my job. That is a clear sign that there was not forethought of how this helps enable success. It was someone decided they needed it.

noticeable thing. Like we spend less on power, we spend less on this. To me, the hard thing about data is that it's just this like, it's floating everywhere. Everything's a data point, right? So when you have a data strategy, what does that mean? And I guess I'm curious, like, on your engagements that you're working with, after you kind of have acknowledged or an organization says, Yeah, I want to work with Centric on building this data strategy. And you've come up with these use cases.

And I assume you kind of know where there's ROI to be had and things like that. What is it that you're actually doing in those hours spent outside of the interviews and building like the strategy, if you will? What is the end result? You know what I mean? Like, what's the tangible thing that like that CFO would then say, Okay, we've got this.

that are drive adoption. You'll see information around, you will see those transitional architectural states, but it's only there to help identify, this is why we believe the investment is going to be as follows because these systems will be able to be deprecated, after a certain point, those sorts of things.

prioritization process. And what they came up with was, okay, I could invest three months and implementing this part of the platform, which was going to be basically getting down to the policy profitability, which is pretty good thing to be able to get insurance. And you'd be surprised how few insurance companies actually know that.

Or, I could spend three months with you guys on building out a complete Analytics Suite around policy submissions and policy quotes, same amount of time, the fundamental difference was, while understanding that profit information was of value. It's something they could kind of do today, it was hard, it was very manual, and it wasn't as granular and the benefits we gave you that detail were reasonable, right? It was more cost savings than anything else.

On the submission and quote stuff, which is basically them looking at real time activity of who's coming to us looking for us to underwrite a policy, and what are the outcomes? Where are, you know, where are we disqualifying them? Where are we saying, yes, we want this business but we're for some reason not getting it. They had no insight in any of that right

now. And they fundamentally knew that if they did, they could get much more lift, they could start targeting policies a lot earlier in the process, where they knew they had a compelling story, they had a compelling rate, etc. And so they opted to go that

route. And the benefits have been massive for them, they now have insight to data that in their 100 year history they've never been able to see before, and their entire sales team is well, that's not an exaggeration to say slightly flipping out for it, because they now have daily data of I now know how to prioritize my day, I know what agencies I want to call and their hit rates are going up.

That's it's a tangible benefit, same amount of time, same amount of investment, but now they're registering top line growth is result of that investment rather than long term potentially being able to realize some cost savings from the other.

matter. And it drives a lot of this, but without the business in mind, and what's your going to do and in that case, you know, you're making it easier for the insurance people to do their jobs. The ROI on it is it's a business outcome. It's better experience. And it's driving, like behaviors for 5 to 10 years. And this is kind of we're talking about the data strategy. So for those of you who have not done this or haven't been that involved, like it's important to understand

kind of the basics of it. So Kris, I want to flip the script a little bit and talk about what everyone wants to talk about, which is AI now. And you can't turn the TV on, AI is everywhere. AI is in some worlds doing people's jobs in some worlds so far from it. And so let's talk about like AI and data. How do they correlate? What does it mean? What are you seeing from your perspective, like for AI implementations? I know from our perspective, you

can't have AI without data. And that's the common thing that people miss. But like, What is your perspective on where we're at with AI? And where data strategies are kind of aligning?

The only reason in the totality it's not is because part of your data strategy should be how are you going to handle data governance, right, you shouldn't, standing up a Data Governance Program entirely by its own without a better understanding of how it impacts your strategy is a recipe for disaster. But the reason governance becomes so important, and partners heavily with strategy is there's really two

components. Everything that everybody sees these days, that is really exciting the world about AI is really one small sliver of AI. It's large language models. Which I mean, it's a combination of a couple of different disciplines within AI but it's still a small part of the overall umbrella that AI

covers. The difference is is this finally is hitting in a way that everybody, even folks that didn't used to really care about tech, it can introduce itself into their lives in ways that are just, feel a bit Star Trek-y, you know what I mean? It's like I, I didn't think that

would ever really happen. But now, you can see how someone in my lifetime probably in the next 12 months, is going to create a prototype as a lark of the computer on the Enterprise using speech to text, text to speech and an LLM.

exist in the cloud. You're not going to get an LLM system that you can put on premise that is of any real value, that you're not going to pay a massive arm and a leg for the hardware for. Two, once you get that information into the cloud, if your data is not clean, understood and secure, the outcomes are going to be catastrophic. So, you know, something folks heard a lot of when ChatGPT first came out is a lot of folks in the news, were talking about how it likes to

hallucinate. Which, in the later versions, they've kind of figured some of that stuff out. But the reality is, is it wasn't hallucinating. It was misinterpreting data, both the question you were asking and the underlying data that was reading the answer. You start introducing a system like that into your environment and you know, we are a huge partner with Microsoft, we do tons of work with them, big fans, and they have inside of Copilot, their

architecture is really cool. And part of what that architecture enables is for Copilot to read any data in your tenant. And it does that so when you ask Copilot a question, Copilot has context with which to answer it,

right. But they also stress extremely heavily, and as partners of theirs, we do the same with our clients that you need to be ready with strong security and strong governance in place before you turn something like that on because the best example I can give, which I'm sure keeps CHROs and CFOs up at night is great, you turned it on and very quickly found out the hard way that somebody put a spreadsheet out

in a share somewhere. They didn't secure it properly, but it was never really a concern because nobody knows that share exists except that person. And in that spreadsheet is a list of all the employees in the company, and their salaries, and social security numbers, etc., relevant data they needed for their job and they didn't realize where they had it, it wasn't necessarily as secure as they wanted. You know, OpenAI doesn't know that. It only knows what you tell it and your

security model. So it's going to see that file, it's going to consume it, put it in this beautiful graph database so it can search against it when you ask it a question. And the first employee that comes along and says, asks your internal OpenAI service the question of, what should I be making as a SQL DBA, ChatGBT is going to see that file and if it sees a role in there called SQL DBA, it's going to start spitting back numbers, because you haven't told it that that person shouldn't have

access to that file. And suddenly, you're going to have information in the hands of people that you never want it in their hands.

really good at it. They struggle to understand like, what are the security risks, like how does it actually affect me and that's a really good example of like, Hey, you keep your files in a place you've always kept your files, and now you enable this tool that you've heard is so great online, across your organization and some IT person convinced you that like you should use it. Now all of a sudden you have this risk that you didn't even know before and now all of a sudden you're

you're reeling. And it's it's quick like the how fast that can happen is so real. So I love the example you gave, and for anyone listening, that should hopefully help you understand like how quickly this can impact your organization, especially with Copilot and others. I want to ask you another broader

question. You know, like over the last, we'll call it 5 to 10 years, compliance has become a big, you know, hot topic here, like you have GDPR and CCPA and I'm not going to define the laws go look them up for anyone listening, but how do you navigate those laws? And are the laws keeping up with what's happening in data and AI, and all things between?

GDPR. Not intentionally, they had not adequately managed their own systems in a way that allowed them to be compliant with their own law. And one of the things that it exposed was in practice, in principle GDPR is a great and necessary law. It's something that as a person who has personal data that somehow manages to get places that I've never even talked to, I'd love the idea of I have the right to tell you, you have no right to that, practically, how easy is it to make those things

real? And what is the burden you're placing on companies when you say you have to do that? Right. Again, not to belabor the world of insurance companies, but some of the largest insurance companies in the world have dozens of systems that contain client information, policyholder information, some of these systems are still 30 years old. They're running on mainframes, AS/400s, S/390s excetera, systems that were built way before the concept of

GDPR was even thought up. And trying to take 30 year old technology and retrofit it in a way where it can be compliant with these modern day laws is extremely cost prohibitive. And it raises a question of, okay, what's the benefit that that law is bringing? And what is the pain that it is exhibiting, it is exerting on businesses who have to comply with it. And as a consumer, I see the upside of, I get to tell them to get rid of

my data. But am I seeing the downside of how much is that increasing my cost of consuming what they have to provide me? A number I'd love to see is how much has the average insurance policy gone up since the creation of GDPR and CCPA? How much has it gone up as a direct result of needed technology investments and attempts to comply with those laws?

argue maybe. I guess let's talk a little bit, let's go back to the cloud conversation because you mentioned Centric, you guys help with the cloud strategy and the data strategy. Cloud has been this conversation for 10 years that I think data AI is becoming. So what does it mean now to have you know, like a data strategy? How does it align with a cloud strategy? What do you, what do you advise a customer on who's kind of going down this path with both?

immensely cost prohibitive. If you didn't have millions of dollars to spend on the hardware, you weren't getting that functionality. And what that meant was 15 years ago, the vast majority of carriers really didn't have data solutions. The big ones started to and they started to invest, they could afford it. The smaller ones, you know, tier threes and lowers, they couldn't afford that kind

of stuff. As Cloud has come along, you can build solutions today in a cloud environment that are as capable as what those massive tier one carriers spent 10s of millions of dollars on 15 years ago, for a 10th or less of that price. And if you don't want to do it in the cloud, you're going to spend a whole lot more than that, and you're going to lose

functionality. You're gonna lose the ability to scale on demand, which is just a massive capability of the platform as a service brings you, you're gonna lose the ability to get access to all of the latest tools that are coming out, you know, OpenAI is an example, right? If you go to Microsoft, they have AOAI, Azure OpenAI. You can't get that

outside of Azure. You want to get scalable systems, right, you this used to happen all the time, it would be when you're building your data center, you had to buy 10 to 20% more hardware than you needed right now, because you needed the ability to scale as you needed it. You don't need to do that in the cloud, you go in, you set up your system, you set it up at a certain level based on your current poll and the amount of concurrency you have in

consumption. And you could actually automate it to monitor, monitoring, excuse me monitor this level of consumption, and in real time, scale up and acquire more resources until utilization drops, and then scale back down. Yeah. And that means that whereas before that would have required you to upfront buy hardware, that would be sitting idle for a long period of time. Now, it's no, once a month, for about four hours after year-end close, everything gets hit, like because everybody wants to know

where the numbers are at. So for four hours a month, I need twice as much hardware. Do I really want to buy twice as much hardware for four hours a month? No, I'm effectively leasing it for those four hours, and then I get it back.

a lot of factors. But COVID, spread everyone out, the amount of data that's being captured is increasing so fast that you'd be buying hardware every day, if you if you really want to keep up and then you know, the larger the organization, the more procurement processes you have to go through. So it's just way easier to scale. And then you start looking at the costs of scaling, increase so much that you actually it's soft costs.

And that's the evaluation we've been looking at over the last, you know, 5, 10 years as AI has kind of become a thing. And now it's here, and it's like, holy crap, this, this data footprint is getting so massive, so quickly. All right, we need to scale with it. So I agree with

you. We see that commonly. But before I get to our kind of our keystone question, we always ask all of our guests to wrap up, I'm curious, in your perspective, you know, your Data Analytics Practice Lead, what a you know, obviously AI is here, right? And we're talking about it in every form of life. But what what specific emerging tech

or tools are you seeing in the data analytics space that you are particularly excited about? Doesn't have to be a specific tool, but it can be a group of tools, or whatever it may be.

in Excel and do filtering. More modern days, it's I've got access to a series of dashboards and I got to figure out, are any of them capable of answering that question? And if so how do I have to slice and dice things. And inevitably, the better you get at it, the more you realize that I've only been able to go so far without being able to customize this thing, like on my own. And not everybody's going to have that skill set nor do you want everybody to have to

have that skill set. And something that I know Microsoft is working on right now, as I'm sure their competitors are, as well. But we do tons of work in PowerBI. It's a really cool tool, right? And one of the downsides is exactly that. If you don't know how to use PowerBI, you're not gonna be able to customize things if your questions start getting more granular beyond what the dashboard was designed to

answer. Well, they're working on a Copilot for PowerBI, where you will fundamentally have the ability to just say, please answer me this question. And Copilot will have the ability to look at everything that your PowerBI environment has access to, and formulate an answer. And that's fundamentally scary that, okay, how are you going to know you can trust the answer versus maybe I misunderstood you or doesn't exactly understand the use of that particular field?

That's where we're starting. And that's something they're going to have out within the next six months. Two years from now, just being able to go to clients and say, yeah, not only are we able to build you this solution, you're gonna get access to this tool where your most inexperienced user is just gonna be able to come in and type in a question like they're putting into Google search and a very reasonable analytical answer is going to come back to them in the form of charts and graphs to

help answer the question. That is empowering people with data in a way where throughout the day in their job, they can ask questions, get real time answers, and be able to impact decisions that they're making. And that's the ultimate goal of anybody that does what we do is empower people to make better decisions with data. Right?

are standing in front of the United States, and they're all watching Kris Moniz, or anywhere in the world. And everyone's watching you and they want to hear what you have to say. You're giving them advice. I heard some earlier but I'm curious, like, what advice do you give if you have one, one piece of advice to give to folks?

yet. And anytime you have an opportunity to make a decision on should I be retaining that data, or letting it disappear into the ether? The answer needs to always be retain it. For the least amount of cost and effort certainly. Find ways to do that in a repetitive fashion where no piece of data goes unstored somewhere. But keep it because you have no idea three years from now, if that data suddenly becomes the difference between you growing 5% or 50%.

favorite platform is. Have a great end to 2023, we will see you in 2024, and thank you for for listening to the IT Matters podcast.