Alrighty. Let's get going. We've got a lot of content to get through today. So if you do have any questions, chuck them into the webinar chat for me. I'll see if I can save some time at the end to answer any questions you might have. Again, this is chock full of good stuff. I think I've got 46 slides, so I kinda went a little bit crazy. You guys are so excited to talk about this topic. But, again, we'll do our best. Today, we're gonna be talking about surviving and thriving in the age of AI disruption. So we're gonna be talking about three key principles that I would recommend every organization make the center of their data and AI strategy. And what do I mean by an age of disruption? I mean that AI, whether you're looking at it by use case or by vertical, has the potential to disrupt and change everything. And so the way you think about data, the way you think about other assets, the way you think about human resourcing, everything, and to the types of use cases you can start to solve, all of that stuff is being completely upended. So let's get going. So I'm Robert Curtis, managing director of Asia Pacific for InterWorks. I've been with InterWorks for about twenty years, so I've been here for a very long time. I've had a great opportunity of working with a lot of customers. I recognize some faces in the chat, so good to see you again. Always happy to see new folks, new faces, new names. Looking forward to touching base with you again very soon. A little bit about InterWorks. We basically do three things. We do data strategy solutions and support, and that's a very simple way of defining a a bespoke solution set of a lot of different things that we can help with. An easier way or more detailed way to think about the broader things we do is this. So starting with strategy, where do you wanna go? What risks do you have? Frameworks you need to build? How you think about people and and technology and the processes that connect them to value? Once you get past that foundational road map plans, long term considerations, then we get down to foundations. And that could be building your platform, installing applications, setting you up for the cloud, whatever your preference might be, implementing governance so that it's not just a theory. It's not just a plan. It's automated. It is predictable. It is, secure. Then taking your data and building it so that it's an asset, building it through the different layers so that you have usable data that can drive business decisions. And that's when we get to solving problems. And, again, analytics, advanced analytics, machine learning, AI, data science, the whole gamut of what's now available to you. Talk to data, all of these things, we are certainly excited about doing. But we don't stop there. Because whenever you think about, a successful project, it's really multiple elements. Right? It's its purpose, which you think of that as strategy, its platforms, which we've talked about foundations, its process, but it's also people. And so we can help support you. We can support your community by helping build that data culture, implement change management, raise your, your data literacy, and we can do that through a number of different ways through recurring or onetime or assessments and accelerators. We also support the individual through training across all the different technologies and partnerships that we support. And if you don't have the appetite, and quite frankly, wouldn't blame you for supporting some of these tools or some of the abilities within these tools, like a data pipeline, for instance, we're happy to support that for you as well. I'm hearing from the chat that there is a bad echo. Can you guys confirm if that's me or that might just be the individual user? Because I've been on my headset all day, and I have not had any other challenges with other folks. Anybody else getting an echo? Just put into the chat? Okay. It might just be your audio. So, yeah, I would I would check on your your settings there, but thank you for pointing it out. Happy to help if it's something that I'm doing wrong. Cool. Let's keep going. So a little bit more about InterWorks. We've been in... In fact, this month, I've officially marked our thirty years of experience from 1996 to today. Our Australia Asia Pacific business has almost been ten years. We've had the ability to work with thousands of customers, some of them the biggest names in the industry. But, again, we're looking for the right relationships so we don't discriminate. We work with all verticals, with all sorts of customers. We're just looking for people that want long term partnerships and genuinely value thought leadership, and, working in close collaboration. I would highly recommend if you find this webinar interesting, go look at our website, interworks.com. We have a blog on there. We're probably more towards 4,000,000 page views, nowadays, with white papers, blog articles, videos, all kinds of stuff on there that's just super dedicated, useful for anything related to data, analytics, governance, AI platforms, RBAC, you name it. We've got it. Alright. That's enough of the preamble. I think I have stalled enough to get a lot of people into our chat while introducing a little bit about who we are and who I am. Let's get started. So, again, I mentioned three really, really important principles so that you can navigate your way through the age of AI disruption without stepping on a landmine or not making a big mistake or potentially missing out because maybe caution rule today a bit too much. So rule number one, own your business logic. I've mentioned this in almost every webinar I've done for the last two years. So the question you need to ask yourself is how portable is your business logic today? So what is business logic? So if you think about all the raw data that you have, it might be sitting in a lake. It might be dispersed all over the place. There might be a bunch of stuff that you have in SharePoint, all this unstructured data, etcetera, etcetera. That's useful, but it's your business logic, the way that you're leveraging it, which is really valuable both to you and to the companies out there that would leverage it for you or leverage it in spite of you. So those are your definitions, calculations, rules, relationship, all the context that you have about how you use your data to make good decisions, to service your customers, service your patients, whatever business you might be in. And on the back of that business logic, that's how you make really impactful decisions. Where do we wanna sell? Who do we wanna sell to? How do we deliver those orders? How do we maximize our profit? How do we raise our delivery standard? All of those things come from your business logic. There was a saying that data is the new gold. It's business logic. It's the semantic layer that is really the most valuable thing that you have and the decisions you make on how you bring all of that stuff together. So logic is your edge. And what do I mean by that? Well, anyone can buy a tendency on Amazon or Azure. Anybody can buy Databricks. Anybody can buy Snowflake. They can get access to all the models, all the BI tools, all the third party data feeds. That's available to anybody. That's not a barrier to entry. What is a barrier to entry is the stuff that only you have. Your relationship with your customers, your relationship with your suppliers, your understanding of your market, and how that translates into your business logic that sits on top of your data. It is the differentiation of what makes you useful, valuable, profitable. As I would say... And and this might sound like hyperbole, but I don't think it is. Your data and the way that you define it within the context of your business is the number one asset that you've got today. You should keep it that way. Make sure you have complete control of how and where and when you use your data. There are a lot of folks that rely fully on third party solutions for the convenience. And as soon as they start to look outside of the prepackaged service that they bought to do anything else, they realize they do not own their logic. They do not own their decisions. They do not own the future. So I wanna help caution you against that. So these are things that have probably been, shared at nausea, but it's good to sort of establish a baseline for us. When you don't have shared definitions, which means you don't have a universal source of truth, a repository upon which all of your business logic is centralized, and you rely heavily on fairly siloed SaaS applications, you will find... And I've been in boardroom meetings with clients where these exact things have been addressed, numbers that look like this. Finance has a sales number or revenue number. Sales has their revenue number, and whatever AI or other tools you have comes up with something completely different. And it's because, largely, these are siloed. Finance has their tool. Sales is using Salesforce. AI, maybe it's ChatGPT, maybe it's Anthropic, but no one has a shared source of truth. Which means when you have conflicting answers, it it it means nobody trusts any of these numbers. I mean, finance probably has the largest stick, but you're paying salespeople off of their number, not finance's number most likely. There is time wasted in terms of the inefficiency of trying to figure out what decision to make because these are fairly large swings. You have to rebuild. You have to calibrate. There's falling trust and decision or your data everywhere, not just in these metrics. And, again, this is because you do not have a centralized repository. You're using specialized tools to do this. With cloud based data warehousing and in particular now in the age of AI, very siloed ERP CRM tools is very much a thing that is fading, if not completely redundant. The other thing that I think is really important is to think about who is consuming your data. AI or agents, are going to be the primary consumers if they're not already. So when you build a dashboard, in the old days, you had semantic layer. You build your dashboard. A person interacts with it, and then make a decision and go do something somewhere else outside of the dashboard to draw value off of the insights that we're getting. That is already redundant and is only going to get more so as we go further. This is more of what the modern workflow looks like or will look like for you depending on where you are in your adoption curve. That semantic layer is now super critical. There used to be that you could have business logic inside your dashboard, and together, they kinda paint a picture that a human being, could act on. It is now critical that everything be in a consolidated semantic layer, a a a well curated data product because all of these different AI agents are gonna be looking at it. And and not just looking at it, and we're giving answers to human users, but oftentimes acting autonomously on it. So you've got to keep that firmly in mind. So if your data is locked in, how many agents can access it? Do the agents have everything they need? Because agents can't sense check. They can't look at a number and say, it doesn't look right. It doesn't pass the smell test. An agent just does stuff. Very powerful, but the higher level of risk than just a dashboard. If there are mistakes in your data, then that scales quite quickly. And if you are thinking in the long term, you have to change paradigm of how you're doing your data engineering. You cannot have a small data engineering team build bronze or silver layer assets and then expect your hoard of data analysts to make sense of it by trapping all the logic inside of your dashboards designed for agents. And, again, we've got a whole bunch of webinars and a white paper on the preeminence of the semantic layer of the age of AI. I'd go to interworks.com if you wanna learn more about that. So just as a recapitalization of everything we just talked about, one governition... One governed definition feeds every tool, supplies all the use cases. And those use cases might be dashboards across multiple tools, or maybe you're using them in spreadsheets like Google Sheets, or you're pulling them down to Excel to do other analyses. I would say look at Sigma if you're doing that. Embedded apps, data apps, AI apps, agents themselves, notebooks that you're doing data science or workbench style analysis on, workflows, all of those things benefit from having a strong central definition. And you need to own those definitions because if they're in, say, SAP or other monolithic old style application, there are barriers for that access. We'll talk more about that in a second. So if you are wanting to own your own business logic, I would say let's start here. Pick the 10, 20, however most, the sort of tip of the iceberg would be for you, which is most critical to your business. It's most useful if there's a lot of disagreement on it because that's the fault line. That's the point of contention. For instance, revenue is a great one. Finance revenue, sales revenue. What do we report to the street, and what do we pay bonuses on? Agree and assign. Build a definition and an owner for each. This is part of the one zero one of data governance. There's the technical data owner, and then there's your data stewards. It's on the business side. There's a lot of different ways you can cut and paste that. If you cannot come to an agreement, give each a unique name and assign when and where and how they're to be used as part of your ontology or data definitions or data dictionary. Put it into code, preferably into a semantic layer that is centralized and extensible. And I would highly recommend that it is SQL based in version control. You do not, as much as you can, avoid proprietary languages that come with a vendor's tool. That might be convenient, but if you're thinking of the long term, it's highly inconvenient. I'll give you some very concrete examples of how in just a second. Open it up. Get value of it through BI tools. We'll talk about what MCP means in a second, but that's the gateway for a lot of different AI to start to use it. And as you go, don't try to boil the ocean. I think it's easiest to start with one domain at a time. I would do that pretty much across all the different sorts of use cases you might solve. Here, we're talking about centralizing your business metrics, but it also might be building AI use cases. And and the very last one that we did the webinar, I think it was scaling AI. I said start with the business domain, prioritize your use cases, and start with the easiest ones that are gonna offer the most value. Same thing here. One domain at a time. So if the metrics that are most important are, say, finance, which is generally where people start, maybe it's supply chain or product. Maybe it's HR. Maybe it's marketing. You guys need to decide. Obviously, consultants like us that have a lot of experience can help you from a strategic standpoint, identifying those use cases and helping you prioritize. Rule number two, get value today. Yes. It is a market of disruption. There are going to be, leaders that change day to day. There are gonna be, models that are released at a increasing frequency, and I'll show you exactly what I mean later on in this presentation. But I don't want you to be so intimidated that you don't put your foot in the water. The only way you're gonna learn the value of AI as well as how to avoid the mistakes is getting your hands dirty. So let's get value today. And I'm gonna give you some strategy guidelines specific to this point that'll help you. Again, own the foundation, find one value... Find value one use case at a time, Date your providers. Don't put a ring on the finger, which, I mean, don't sign an all inclusive multiyear agreement, particularly from an AI perspective because, again, there is a lot of change that's going to happen. For example, well, I'll I'll I'll tell you that story when we get to that point. Governance needs to be foundational and pervasive, so it needs to travel wherever your data goes versus being coded at the destination. And and, for always, keep testing the accuracy and usefulness of the model, but also the ROI that you're getting. And if you need to pivot because you are light touch and you're thinking of this in a value today versus long term, let's go all in strategy in a very volatile market, you have the ability to pivot. So let's get into each of these in a little bit more detail. So, again, own the foundation. You marry your cloud provider. You marry your cloud based data warehouse. You don't need to marry your AI. Now there's a lot of cool stuff that Snowflake or Databricks or whatever have inside of it. But the thing that's really valuable about those is that they have the ability for you to go grab AI models and agents off the marketplace, put it into a container, and run it inside of your environment. You're not marrying, these, these AI providers. There's a lot of value, and there's clearly some exciting stuff happening, but everything is at risk. It's still at the minute. There's not a settling, in my opinion. So, obviously, yours to keep. Your business logic, we just talked about that. Make it SQL based. Make it portable. Do your best to do the hard work of coming to shared definitions. And when you're thinking about a data platform, you really want one that has extensibility and interoperability. So open formats like Apache Iceberg, for instance, or or or those that are really embracing the broadest use of MCP, which we'll talk about a little bit later. Commit to the platform for the long term and keep the assets that you need portable. Platform... A long term platform does not mean that you lock it into the proprietary language. Again, remember, keep it SQL. Find value one use case at a time. Pick a problem. Prove it fast. Fail quick. You don't wanna throw years worth of it to find at the end of it. And then what works, scale, production wise, and then move on. Repeat. Repeat. Repeat. Hopefully, for those folks that have been working on data and building pipelines, I have a sort of rapid prototyping agile approach is something you're familiar Flexibility is key. I did a quick search, and there's 97 different AI providers, probably more. And across those, there's almost 1,400 models. ChatGPT was seen as untouchable in 2025, but then Anthropic exploded ahead with unprecedented growth. There will continue to be market fluctuations with new leaders. There may be an AI bubble, and I'll talk to you about that here in a second, as well as emerging technology. LLM has changed everything, but is there something else on the way that'll completely revolutionize the way we think about AI again? Potentially. Anthropic is putting together a... An I... An IPO. And so there's information out there, and I've read some of the analysis. And it seems like I can't miss. But the reason I'm saying, for instance, across all of AI, I wouldn't put a ring on a finger just yet, is even as untouchable as Anthropic looks this year as opposed to chat chat GPT last year, is when you actually read the briefing, Anthropic has a awesome amount of revenue. But when you look at profit, they're losing money hand over fist, and it's because they're paying so much money for these compute farms, for these server farms. And they are obligated to pay those fees for those server farms years, a decade into the future for... Based off of their contracts. So if Anthropic's revenue doesn't massively continue to grow, they're at risk. And so some other model, some other provider might then step over Anthropic. So that's why I'm saying it is an age of disruption, not only the things that's AI making possible, but who actually settles into the top one, two, or three leaders. That's why people think there may be a bubble. So that's why I'm urging you to not get caught in that. Stay light. Stay agile. Make governance travel with the data. So whatever it is that you're using, Copilot, Claude, whatever, have them go through your governance layer that's inside of your warehouse. Don't think that the governance should be, there at the destination. That way, if another model shows up that's useful or there's a radical shift across multiple things, you're not stuck having to rebuild and replicate things that you've already built. Foundational is what you should be thinking. And then the AI, that's where you do your experimentations and POCs. And keep testing. Evaluate your own data. Evaluate the data quality. Evaluate the types of questions that you're answering and the answers that you're getting back. As you learn, it informs you and guides you on how to choose for the next one or to modify and adapt on the existing ones. Deploy, particularly around open interfaces that give you control. That doesn't mean that they necessarily have to be outside of your environment. Your data has to travel to them. Platforms like Snowflake and, Databricks and other things allow those models to come into you or rather to come into your environment. And, of course, monitor and just repeat, repeat, repeat, repeat. One thing that you hear is horror stories. As people roll out a model or they'll roll out claw to a whole bunch of people, and then they'll see how much did those tokens actually cost without guidance, without governance, without a strong control on cost. So monitoring is not something you do at the end. It's not something you do periodically. It is something you actively do. Super important, particularly in the age of consumption based pricing. We're on to rule number three. Avoid lock in. In a highly dynamic market, flexibility is essential. So the risks of vendor lock in are numerous, but here's here's a really compelling four, I think. Number one, if you pick something that is all in in terms of how you code and store your data and use your business logic, You only have options to the AI solutions that are out on the market as opposed to what's in their road map and how they wanna open it up. Vendors do and often to change the rules quite quickly. I come from a very long Tableau background. My company, InterWorks was the very first Tableau partner, first partner of the year, first gold partner, you name it. We were the very first Salesforce bought Tableau. And if you're the Salesforce stack, that's probably a great thing for you. If you weren't in the Salesforce stack, well, a lot of things changed very quickly, and it opened the door for a lot of other analytics tools to start compete for that business. So things do change a a 100%. The trap is this, and you'll see a little bit more about this later. It is easy. It's like a shark's mouth. And I don't mean to be cynical about every vendor, but we're talking about the business model of the of the tool. And every vendor wants you to have easy access into their tools, but they want it to be very hard to get out. This is why companies like SAP and Salesforce have the highest customer retention rates in the world because it is so hard to get off of their platform even if you wanted to, which gets you to the cost of leaving. And I'll I'll give you some averages here in a second. But once you are in in a vendor that is highly proprietary, very hard to get out, the cost of leaving is years of effort. And I have done InterWorks, and and my team in... Specifically in Australia have done those exact projects to help people get off of SAP or Teradata or Redshift or all of these other tools where it is much harder to leave, particularly things that have a a particular blend that grew into other things like an ERP that is wanting to be your data repository or CRM. Put all of your data with us. So what does vendor lock lock in look like? So it might be a proprietary language, like, say, LookML inside of Google Looker. It might be closed formats. You have to comply to our way of doing things. It also might be no open interfaces. You... There is an API, and it's very limited, or there isn't one, and you have to do exports. All of these things in single... As a single problem or altogether is where... How your business logic gets trapped, which means if you're going to get it out, it is a very manual or arduous process to do it. And let me reiterate why it's important to have this flexibility. This is the release of your top six AI providers, by half year. So in 2024 from January to June, '16, then '18, the next six months, '22, '21, '21, '28. And we're not even through the second half of twenty twenty six yet. These models are not only releasing faster, they're getting step change, magnitudes better. And I think anybody that's in this industry can say the models from six months ago don't have a candle to the ones today. A 126 new models since January 2024. You've got to plan for adjustment, adaptation, pivoting. If something really game changing comes out, that's gonna really help your business. If you are locked in and your provider hasn't put that on the road map or sees it as a comp... Competitor, you might be stuck. The average gap in a release is every eight days. That's down to six days just in 2026. So, again, this this whole AI wave is accelerating. The value is increasing. So it plays... It makes the most sense to be adaptable, to float, and find wins where you can. And then once it settles, then look at bigger commitments. I mentioned MCP a few times, so I'll just define it. Model context protocol, it was introduced by Anthropic, about two years ago, actually, almost exactly two years ago as an open standard than other tools, OpenAI, Gemini, etcetera, agreed to open it. But that doesn't mean MCP has now leveled it, and anywhere you wanna put your data, you can. MCP addresses access, not ownership. The vendors control how they implement it. And oftentimes, AI models going through that doorway talk to your vendors' agents, not to your data, which then obviously puts that middle person in, you know, a a middle object in in between the the value that you're getting. You you might be paying twice for both vendors, for both agents doing things. And, also, how good is that agent in in relaying exactly what you wanna do? Now there are plans. Like, for instance, I don't think Databricks allows you to have direct access to the... You you go through, Genie spaces, but it's on the road map to to fix So depending on where you are, which tool you're on, or what plans you have, it's really worth starting to think how close is the system. And I've got an analysis that I did. I was gonna share some of it. I was like, I don't wanna point fingers at anybody directly in a public forum. So if you want advice or opinions on Fabric, Databricks, Snowflake, Google, or Salesforce, reach out. I'm happy to show you my opinions in a nonpublic form out of respect for, some of the people we still work with. Five lists of rock... I'm... Today is, terrible. Five risks of lock in. So lost leverage, switching costs, I e, migrating, exceed any price increase. So for instance, this is how I see a lot of these companies do it. We want you to join us, and we've got a full stack, and everything you could ever want is over here. Isn't that amazing? We've got AI. We've got data transformation. We've got analytics. We've got a data warehouse, and we've got cold storage like a lake. Come over here. You'll love it. Wow. That's expensive. But we'll give you a 95% discount. Okay. Great. We'll even lock it in for two years. Oh, amazing. So I can't lose. You do the work. You get all your stuff over, and every piece of logic that you recode into their proprietary language is more and more and more of those sharp teeth pointing back. And, inevitably... And I'm telling you, I have seen this directly. I have talked to customers that are in midst of trying to mitigate this. At that two year period or whatever the end of the term is, the salespeople get incentivized off of growth. And if you're going to maintain, you're not going to get that discount. In fact, it is part of the strategy for them to maximize the earning for them to remove that discount. If not wholly, then impart. Because they know once you've got this adoption and you've got every... All of your business logic and all these reports and everything doing in their platform, them adding another 500 k onto your licensing is still less than you having to migrate out. And that's where the starts to happen for clients. There's a lot here, but the only other one I think is worth sort of calling out, maybe the bottom two there, is your provider that you have chosen or thinking about choosing may have pretrained agents ready to go. And they might be really cool, they might be really effective, but you're now on their R and D plan. And, again, if they haven't opened the door fully to the models you need through that, MCP or however, you plan on accessing those other models, then you are very much waiting for them to release or update or give you the options that you might... If you had the full market to choose from. So there is value in those pretrained AI bots, but that's... That value is becoming less as that stuff becomes easier for you to do for yourself. The other thing is is there is a risk, that you have based off of the decisions that organizations make for their future. Australia is an island in the middle of the ocean, and so we have very specific data sovereignty laws. There are also additional laws that we put on people whether they are in particular industries like, wagering or maybe you have a foreign owner, so you've got firm compliance. Maybe you have medical data. Maybe you're government. All of these things become part of a holistic package that you have to consider. The the vendor obviously, did some level of compliance assurance when you started. But if their market shifts and they're like, we don't wanna be in government anymore, or we don't... We... We're not we're not we're not worried about data sovereignty because all of our data centers are based out of this country. Whatever it might be, you inherit that risk. And the harder it is for you to migrate, the more urgency... Let me give you another example. If you have to make a change quickly, there's a paradigm, a common saying in consulting. There's three aspects that you can sort of choose from. You can have something that's great, that's quick, or you can have something that is cheap, less expensive. And in the world of consulting, we tell you you get to choose two. If you want something that's great and quick, it will not be cheap. If you want something that is, quick and cheap, it will not be great. It's impossible. So if you are not at the... If you're not in control of your destiny when you make your migration or the ease at which you're able to do that shift to a different platform because the vendor has changed the product strategy, I've seen organizations in a real crisis as they sort of say, how do we mitigate this? The renewal squeeze. So I just talked about this, but let's specifically talk about it. Year one, you get sold the dream, the art of the possible. And, again, I'm not saying these platforms don't offer value, but they know how to cover up the fact that you're gonna be in for a long time. And if that's great for you, great. I'm not saying don't pick any one of these tools if it's... If it fits your long term solution. But if you're interested in following the wave of AI and all the new things that can happen, I would recommend keeping this firmly in mind. So getting in, you probably have vendor people, presales people, delivery people, and they're helping you guys get in. We'll give you training. We'll build use cases for you because they want the logic. They want your use cases in their platform. Learning curve, the barrier might be high, but you're you're a lot of assistance, a lot of hands on help. Year two and three, that's the deep adoption. That's where your business translation gets into the new platform. You got use cases happening everywhere. And now you could imagine it's at that point where your renewal becomes a real point of contention. Either grow at the rate that you guys bought from us, and we'll keep those discounts because the salesperson gets paid on their targets, and their managers all the way up the line get paid, or those discounts start to go away. And then what you thought was a great deal in spite of maybe losing some versatility instead of, some ability to flow and go to other places because the price was so compelling, that starts to disappear. And then, of course, as I mentioned before, the exit of leaving that platform is so onerous that you're sort of stuck. So, again, I don't wanna belabor the point too much, but here's another slide. In terms of when your logic is trapped, the logic is stopping yours. The metrics drift because, inevitably, you're trying to do stuff, and maybe the platform you've fully invested in doesn't do anything. So other shadow tools start popping up, and then the metrics start to diverge because, obviously, you don't have a centralized place for those metrics to exist upon which multiple tools can draw from. And then the harder and more proprietary it is, the innovation starts to become more slow because that speed of developing, particularly as tools evolve... Just gonna take a quick drink. It means you've got specialists that centralize all of your innovation. Let me give you an example. SAP is a notorious, at least in our organization, for how hard it is to get that information out. I have people reaching out to us all the time about helping them get this out. Sap is a German company. A lot of the documentation and fields are written in German, just as another example. But I went and did a little bit of research on averages. And so migrating even within SAP, you're seeing the average cost 4 to 10,000,000. From our experiences, we have seen those numbers, over multiple years. Fully leaving the SAP ecosystem is a degree more complex. The larger programs, these can get significantly, more expensive. And McKinsey reported, it's very common to have a two to three common budget overrun. And I will tell you, I have seen organizations that thought they were setting up for a 20,000,000 migration... $20,000,000 migration and end up in the hundreds of millions. There are a lot of organizations. There are a lot of tools that have a similar degree of difficulty getting out as SAP. So I would say make sure that you are very aware. Where does that money go? Well, consulting and solution integrator fees. This is a a lot of what we do, helping people get to tools that are more usable, less legacy, more progressive thinking, more integrated with future opportunities with AI. There is a lot of money that has to be paid to dual running because you can't shift... You can't shut one off on day three sixty five, and on day one of the new tool, have the other one just go. You've got to actually co build and get stuff moved over. And the more proprietary is the longer it takes. Six to twelve months is actually fairly conservative. If you've got a fully developed data enterprise or data warehouse at the enterprise level, it could take you years to do that. And it's not just the ERP, such as SAP. Here's some other stats for you. 57% of enterprise platform migrations cost more than a million bucks. 25%... Only 25% achieve the expected value within a year. And the biggest cost is the engineering time. It's translating it from this language to that language, the same business metrics that you that you defined. So how would you make good decisions from the start? Well, I would start like this. I would say, hey. Mister vendor or vendors that are competing for my business, show me how easy it is if I needed to exit your platform. How extensible is my logic? How portable is it? What's your interoperability with other stuff? Can I keep my business logic in something that is very open, like SQL or DBT, YAML, whatever, so that it? From a semantic layer perspective, I can hook numerous tools off of it directly versus having to go through your API. Version control. How do I version test document metric definitions? If I did leave, contractually, what are those things I need to be worried about? Who... I I would... I don't know if you get this from the vendor. You can ask. But I'd go find other people that left the platform and ask, why did you leave, and how hard was it? You you you you probably, depending on the tools you're using, have people in your network that you can go and get fairly unbiased opinions of from a customer perspective. I do regular audits. Have the customer... Or rather have the, the salespeople or your presales people or your customer success people come to you, talk about changes, talk about anything that's gonna affect your security compliance regulatory requirements, and be very strict and be very straight with them. All of these things are important. I'd be remiss if I didn't sort of pause here and say, InterWorks does spec and select for our customers. So if you're thinking about a data warehouse, a governance tool, analytics, AI, whatever, we could run these. And sometimes they're as short as a couple of weeks. Let's get a couple people in. Let's get you demoed. We'll give you some perspectives in terms of how it's gonna relate to your bespoke standards in your data and your use cases. If you wanna go deep, we've done these that have lasted six months or more because you really need to make a smart decision for a long term strategy. So we hope we can take some of the mystery out of this. And speaking of helping, we have several things that we do, but one of them that we're gonna offer to you is called a DART. This is a lighter version of a DART, I still think it's super useful. It stands for a data and analytics review and tactics, specifically focusing on your readiness for AI. To productionalize AI at scale, you really need a good foundation. And so the Dart does an assessment across five key technical domains. That is your culture, I. E, your people. People need to understand how to use AI, both both how to get the most out of it, but also how to use it ethically. Two, your analytics, the types of questions you're able to ask, the frequency in which you can answer them, the quality of the answers that you're getting, the tools that you're using, all of those things fall in that category. Three, potentially, one of the most important is your data. Not so much the data that you have. Obviously, the more data you have, the more answers you can answer. But, really, the the platform and processes, the orchestration, the automation, all of those things that make that data into something usable as a product for analytics or AI or whomever. So we're really looking at the maturity of how you've built a data platform. Number four would be your governance, the rules, the security, privacy, how you classified it. Could you detect whether you've got PII information? What rules do you have around unstructured data? How are you automating these rules so that it's not just human best effort? And then the last framework, the last pillar of that framework rather is platforms. How prepared are you, for recovery? How insulated are you from disaster? What is your cloud to on prem strategy? Etcetera, etcetera, etcetera. We'd love to help. So if you've got an opportunity for us to have a conversation with you, this doesn't take much time. It's five workshops with surveys for each, and then we can get you back a very concrete tangible report as a result of it. It's very visual too. So it... It's useful for folks that are very technical, but it's also useful for your executives. And it's something that you can repeat about every year or so so you see a visible change as you start to make improvements and and follow-up on some of the advice we're giving you. Probably should've gone to this slide to do all that explanation, but there you can see visually culture analytics, data governance, and platforms. So if you need help on any of the things that sort of underpin data, which would be strategy, data, or governance, or you wanna jump straight into how you operationalize and scale AI, You can go to interworks.com. You can respond back to any of the emails we sent with with this or the recording of this email that... Or this this webinar that's gonna come to you, or you can scan here. Obviously, we've got a ton of stuff on interworks.com, which we'd love for you to have a look at, but we're always here to help. Awesome. So we do have some time for questions. I've got two here in the chat already. So please add more questions as I answer these and see if we can give you guys as much extra bits as possible. So the easiest question in here so far is will I be sharing the presentation deck? We'll be sharing the video for sure. That's automatically gonna go up on the website. And if you would like the presentation, just just respond back to the email, any of the emails that we sent you and ask, and I'd be happy to send it over. And if you wanna have a deeper conversation about some of the specific tools and how we would rate them in their portability or openness for your data to go and do other things versus being trapped, I'm happy to have that conversation too. The other question I've got in here, is please explain the inconsistent metrics, five risks of lock in. So I'll go back into our deck, which I think was right, here in their lock in. But to better explain it, I'm gonna go back to this slide right... Trying to do this from right here. So the risk of inconsistent metrics. So if I've got three very hard to use tools... I mean, I mean, hard and getting in there clicking buttons. I mean, hard for the data to move between them. Like, say, for instance, we've got TM one sitting next to Salesforce, and then we've got ChatGPT over here. And ChatGPT is looking at our data lake. It wouldn't be uncommon for you to end up with vastly different numbers like this because finance, owns their database, puts in their specific rules in isolation because they can do that because it is their tool. And, again, that tool is likely to be very hard to get data out of, and finance is happy to use it from that tool versus from a centralized repository. Sales, on the other hand, they're getting their data from some source system, and they're doing what they want with it in Salesforce or other CRM tool. And then the AI is looking at the, the raw data, maybe different databases, maybe ones that have been replicated and and not, repaired as business logic change or or records or fields get updated. So that's the risk there. So if you had, at the very foundation of your data strategy, we're gonna centralize everything. So the bronze layer, everything from these source systems, these databases, these Excel sheets, whatever, from our vendors, suppliers, whatever, through their system, everything gets put into one big spot in Snowflake. Then we build our silver layer on top of it. So let... How do these things all join together? What are the relationships? Let's do some data quality, deduping, etcetera. Then we build gold on top of it, some relationships, some logic, some definitions. You then have finance, sales, and AI, and marketing, and everybody else all looking at the same... Not just same data, but the same rules, definitions. So when we talked about what business logic is, all of those things have been... And governance have been applied to one spot, which means you've got everybody working off of the same numbers because they're working off of the same data, definitions, logic, governance, access, etcetera. So when I say inconsistent metrics, that's what I'm talking about. And the harder it is for your platforms to get data out, the less likely it is. The more effort it's gonna take for you to centralize that. And if you centralize it in a platform like Databricks or Snowflake, we love Snowflake, it means that all of these AI agents, the hundreds that are being released... I wanna say hundreds, but since 2024, a hundred and twenty six or whatever the number was, But but coming out almost weekly, all of the new cool stuff can actually point at the single place where you know you've got the best version of your data, and you get the benefit of the innovation happening all around you. So that's what I mean. Hopefully, that answered your question, Suhon. Do we have any other questions? We've got about ten minutes. If you guys have any other questions, I'm happy to answer. Otherwise, I'm happy to give you ten minutes back to your day. The recording will be posted on to works.com. You will get an email, if you registered, and anybody that registered in, if they didn't attend, to say, hey. Here's the recording. If you would like the deck specifically or to, like, have additional conversations with us, reach out and... To any of our salespeople or to interworks.com slash contact us. Very happy to have a conversation. Awesome. Thank you, Andy. Thank you, everybody that attended. It is certainly a pleasure to have these chats with you, and looking forward to the next one. I think we've got one more webinar for the year, and then we'll give you guys a little bit of a break before we kick off probably February, March. One more to go. Thanks, everybody.
