AI Enabled Mergers and Acquisitions/Planning to Execution at SKILup Days 2024

Hi everybody. Uh, welcome to my session on, uh, generative ai. I'm gonna provide a generative AI primer and a practitioner's guide for AI and mergers and acquisitions.

Before we get into that, maybe I can introduce myself. Hi, I'm Joseph Phoenix. I'm the managing director of In, uh, emerging Technologies and AI at Enterprise Vision Technologies, and I also coordinate the countdown to a GI, which is Artificial General Intelligence Newsletter.

So, a little bit back to the Practitioner's Guide and, and why this is important and why practitioners need to understand ai. And this is right out of the Prince two seven AI Practice Guide. So the guide shares with us that enhanced decision making, along with automation, efficient efficiencies, data-driven insights, skill augmentation, adaptation, and adoption and practical scenarios are very important for practitioners in this modern age of, of ai.

So I thought it would be valuable to share my history of generative ai, which is the latest and greatest craze in artificial intelligence. In addition, on going through a, a, a large project of mergers and acquisitions and stepping through how we might gain value through AI as practitioners so that you can, uh, apply these in your day-to-day practices. But before we do that, I wanna share a little bit about our organization and how we look at, um, these things from an AI perspective.

So at the top, we always have the organization that we support and our customer's why. And then our, we as practitioners want to understand the why of the customer or the person that we're assisting so that we can make that our why, so that we have a, a north star for both us and the organization that we're supporting. And then on top of that, there are various business units and operating units that we may need to work together with cross organizationally to align with that why.

And ultimately, this comes down to enterprise applications in many thing, in many situations. In addition to enterprise support, underlying all of those things, our organization focuses on ITSM, cybersecurity, cloud modern networks, DevOps, and automation. But surrounding all of those things, we see this secure data intelligence, and that's often for AI for the business and AI for your technology professionals.

But all of that is the foundation, the concrete, the cement, the inter of enterprise data and applications that supports these AI initiatives. Now down at the bottom, you can see there's this platform operating model, and we'll get a little bit more into that, what that means, if you haven't heard about it, I'll share a little bit about that. But really it's about synthesizing a, a set of tools and practices that the organization can focus on to accomplish a specific task.

Many large organizations that we work with have what they say, one of everything, but that's not great because what we wanna do is we wanna have our teams become subject matter experts, and that's where the concept of a platform comes on, so that we can have a set of standards that we can apply to these technologies. I'm gonna take a little step back and talk a little bit more about the rapid pace of advancement that's happening in technology. Now on this chart here, you can see at the bottom left hand corner, you can see that in 2020, uh, this company Nvidia, who produces these graphical processing units, which is kind of the underlying machine for ai, they released this a 100.

And that machine had a, uh, this is gonna be technical here, uh, 312 tariff flops of processing power, while only two years later here in 2022, they released this H 100, which was 10 times more powerful and one PETA flop. And just recently, a few weeks ago, um, they released a B 200, which is around nine to 10 PAF flops. So what does that mean?

Why is that so important? Well, every two to two and a half years, this is not growing. One x two x three x larger, this field is growing exponentially every two to two and a half years.

It's becoming 10 XA hundred x, right? A thousand x more, uh, faster every few years. So you can see here over the next five years, we have a projected 100 x growth from AI where it's at today.

So if you're looking at AI with chat GPT or some of these tools, imagine it twice as good. You know, imagine it five times as good. Now if you can imagine it 10 times as good as it is, and then a hundred times, this is why this is important to practitioners to keep up to, um, to date on skills and to really understand the TRA trajectory.

And this is why I curate the countdown to a GII have children. It's important to me to understand what jobs are going to be important in the future and which jobs that we really need to understand AI may disrupt. So with that, I'm gonna share a little bit more of timeline for me personally for generative ai, and then we'll get into mergers and acquisitions and how we can assist, how AI can assist you in mergers and acquisitions.

And I want you to think about how this applies generally, but it's gonna be specific to how we see mergers and acquisitions, but this will be also in general how it will apply to large projects. So stepping back to April of 2023, a good friend of mine, Jean Kim and I, were at, at the Advancing in Women in Technology charity event. And Jean and I sat down after the event and at lunchtime and started talking about the things that we were working on in generative ai.

And that's all we ended up talking about the entire rest of the, the meal. And it was amazing. And Gene had shared me with some things that he had been wanting to do for a long time, and, and we got together and started working on those things.

But one of the other things that we did with the Techstrong group is we created a gen AI for ops architecture back in June of 2023. It was amazing hackathon, and I think it was probably the first ever practitioners that got together from the DevOps and DevSecOps community to talk about how we would apply the principles of, uh, generative AI in a DevOps setting. And this was one of the, the, the artifacts that came out about, about that.

And then Gene and I got together in October, and we built a retrieval, augmented generation chatbot for Gene. Now, gene has about, uh, well over a thousand talks. Now, he can't necessarily watch all thousand of those, but he can have an AI summarize them, and then he can and provide that ai.

So if you go to Gene's site, you'll see his chat bot that provides a summary where you can look up and, and find videos and talks from practitioners in, in, in space and, and in, in large enterprises as well. And if you notice the date here, and back in October of 2023, if you're familiar with what they call the context window of ai, it was at 2000 words. That means that if you put more than 2000 words into your chat, GPT, you'd get this error and it would say something like, oh, tokens exceeded.

So this was early days of DevOps when we were building this. So then we moved on to building, um, more Im, um, impactful. Uh, this is a, a, a tool, uh, for, uh, retrieval augmented generation in the utility space for safety, right, safety guidelines and looking up safety guidelines and validating these safety guidelines for people who are out dealing with electricity and high voltage.

So very important. And whenever they're working out on the field, they have to make sure that they're checking all the boxes and they're up on the latest and greatest best practices for safety. So this was a tool that we had built, um, for those people to leverage.

In May of 2024, we started talking about these coding agents and John Willis and myself, um, published this, um, discussion about code agents. I recommend you take a look at it if you are in the dev coding and development or if you just are interested. Um, the AI actually is very, very good at writing software.

So if you ever had a, um, desire to write software, something for you, or if you're in the AI and technology industry, these sort of chat bots are going to be, um, participating in your projects as AI coding agents. So you're gonna need to know what is going to be issued to the AI to do and what the people are going to do. And this is gonna be, this cooperation between people and AI is going to get tighter and tighter woven into the projects that we are going to be managing.

So it's very important. Um, there's another thing that comes about, and this is one of the reasons why I wanted to share the mergers and acquisitions use case, because many organizations are trying to learn what is real about ai, what can AI actually do versus what it can't do. So we started putting together these concept hackathons where we get together with people, get them excited, share with them what they can do in ai, what's the reality, and then have them take their ideas, uh, for their particular business and then create that into a concept.

And then bringing in AI practitioners to assist them to make sure that those things are within the scope of what AI can do for them. This particular one was actually amazing. We had over a hundred participants, John Willis and myself hosted it, and it was amazing event here in August.

And then just very recently too, you might take a look at, um, the autonomous AI in the enterprise. Now, this was, uh, co-authored by myself and several other esteemed authors here. And this gives, you're kind of living your life vicariously through a large organization that's trying to do a big project.

So if you get time, I recommend taking a look at it. And it really focuses on these things. Cross-functional collaboration, understanding security, first, strategic balance, and really understanding technical debt.

And we'll, we'll get into this technical debt, uh, tsunami that we believe is sort of coming our way with what AI is doing. So again, thank you very much. Now we're gonna go get into the generative AI projects and enterprise.

So with that, if we look at the concepts that we're seeing a lot in, in businesses, and you may hear a lot about these things, these are sort of common practices for large projects that are coming in where we talk about value streams and, and what we see here at the top, this is sort of a traditional plan, build, deliver, run. This is used in your IT for it, your it vil V four also in your architecture like toga and cobit, you'll see these, these, um, concepts of planning to delivery and the value streams in between them. Then if we look underneath that, obviously we have our DevSecOps teams, we have our CMDB, and we have underlying these things, this concept of the SRE.

Now, the SRE is what we call a site reliability engineer. This is a, a practitioner who, that we want to be able to assist us through the planning of our projects all the way through the running and operational deployment of these projects. On the left hand edge here, you can see, uh, it says Google and two methods.

Now, what Google says is what they wanna do is they want to try to get down to two standard methods or sets of tools to do any one thing. Why is that important? Because each time you add a new tool or a new method, you have exponential amounts of increase of resources that have to do this one thing.

That's why our projects, we wanna move towards platforms, as we talked about before, because the platforms allow us to have subject matter expertise and have experts that can focus on these things. So if you get a chance, take a look at site reliability in, uh, engineering. It underpins many of these things that we're doing for projects and really sort of justifies and gives some guideposts on why we wanna reduce the amount of methods that we have to accomplish.

The same thing in, in organizations on the right hand edge, this is taken from what they call the common service data model. If you're leveraging a, uh, service Now or some other tool for ITSM, you'll be aware of this common service data model, but it's a good framework for the way that applications and projects are being managed, uh, today, traditionally, and with AI and AI ops. And the concept here is that we have our business applications sitting at the top.

We have our support for those business applications, uh, which are our application services. We have, again, our platforms and environments, our dev test, QAT production, and obviously we have our data that's underlying all of those things. So if you take a big enterprise application like an Oracle ERP application or an SAP application, that's really the business app, then we wanna break that down so that we can manage that and uplift those applications over time.

And as you can see here, our teams are very important on how we, uh, manage resources with these things. So now stepping back, and we're gonna get right into this AI here. So one of the biggest things that generative AI practitioners, the, the top people when surveyed, uh, said was a problem, was the data.

So you're going to see a lot more focus on organizations that want to uplift their data intelligence and, and what does that mean? That really means that they need to understand how their data is coming into the environment, and then they need to understand the provenance as they take that data, join it with other data sets, create new data from those data sets, and transform that into AI tools. That transparency is really needed for organizations, and it's very, very important for us to really understand those things.

And why is that? Well, one of the big reasons is we need to take advantage of these, this data, but it's also in regulation, the EU guidelines, and then very recently, SB 10 47 in California says that we need to understand the provenance. Where did this data come from?

And when we are going to be audited, we need to understand those things. So as practitioners, you need to understand these concepts and regulations. And if your project is going to be in regulated areas, this is something that you really need to understand as a project management professional.

But underlying the applications themselves, again, you can see the number one challenge here. 70% of participants said data is the number one challenge. Now, if we look over here for the, the other participants, their strategy is going to be their number one challenge.

Now, that's why we say data intelligence is your strategy now. So for us, when we step through the rest of this, uh, use case, I'm going to share with you some, um, exciting things about ai. But before we do that, uh, again, I wanna really just resonate with you on a few things about these projects.

So you're going to hear something called retrieval augmented Generation, if you haven't heard of it. And retrieval augmented generation is a way that AI can look at a, a specific piece of documentation and utilize that documentation as facts. And what does that mean?

The AI has been trained on a lot of data, but that data is only based on the dates associated with its training. So we bring in documents and document knowledge and databases and things of that nature that are up to date information. And what retrieval augmented generation does is it retrieves that information, it augments the context window, and then it generates the latest, greatest information based on the data provided.

And as you can see here, this red line through January of 2022, up into the beginning of 2024, and look at this spike going up, right? This is vector databases. We're going to con continue to see many organizations deploying projects associated with vector databases.

So it's important. And, and why as that, again, you'll hear this term called hallucinations and retrieval. Augmented generation assists us in dealing with hallucinations and can be make our generative AI or our AI subject matter experts on the things that we're working on.

It can even be a subject matter expert on PRINCE two and our project management if we need to reference those things. So with that, I want to take a another step back here and take a look at, there's two major things that are happening, trends that are happening in ai. You're going to see AI that will be on these cloud providers or SaaS models like a chat GPT.

You're also going to see these open source models like Meta's llama model along with the minstrel models. One of the things that we've seen a large adoption of moving into cloud over the last several years, but now with, uh, GPUs and these models, we're starting to see actually a lot of organizations want to focus on building up their data centers again, which is interesting, right? What's, what's new is old.

What's old is new. It's very interesting about what's happening. And one of the reasons is, is once these GPUs start going in the cloud, it's not a scenario typically where we can shut them down, right?

And when we are building a cloud operating model or finops, those things are how do we turn things off and bring things up when we need the compute? Well, if the GPUs are running 24 7 inferencing and serving data or training, our cost profile sometimes is difficult to integrate into our cloud operating models. So you'll see many organizations that were trying to get out of the data center really taking a look at their facilities and saying, actually, it may be more cost effective to run some of these generative AI use cases in our data center.

And I provide with you some economics associated with that. You can see here in purple the GPT-4 and the GPT 30 2K models. And then at the bottom, you can see the cost of running these on some open source models that could potentially run in your data center.

So as the projects come about, you're gonna see more potential hybrid and more infrastructure, whereas primarily, you're probably seeing cloud use cases and projects. You might start seeing more infrastructure related projects depending on where you're focused on in your practice. So I wanna also talk about this, uh, concept of technical debt, right?

Um, we have a lot of AI things coming at us from a lot of different vendors. Now, what is, why is that important? Well, what we want to do is we want to, as practitioners, recommend the right next steps for people to do and assist them in that.

And that's what we'll do with the mergers and acquisitions here in a few minutes. But with that, there's going to be a business case that gets created, right? Right out of the guidebook, our use cases, we're going to pilot those implementations, which are going to have some costs associated with those pilots.

But once we get that pilot in and start industrializing that project, what we want it to do is we want to have this decreasing amount of opex and increasing amount of opex savings and capabilities as we go into full scale rollout of our AI applications. But this reality, we really need to take a look at the, and understanding sort of technical debt. And why is that?

Because sometimes if we don't look at technical debt, right? We'll get into these things. Some of the things like regulation I mentioned, and they can be very troublesome.

So as project management, um, practitioners, it's important for you to know these things because these are risks that you may need to assess. So in this scenario, imagine a project starts again. We have the same rollout, the same cost, and at the beginning of it, it's great we're get getting these cost savings, right?

But something happens regulation, right? We're not sure what the regulation is. Maybe we didn't understand and have the concrete of our data.

Well, that brought up some challenges for us in our organization. Now we've gotta go back and create a stable foundation for our data, because guess what? The regulations came in, but we didn't understand what that was going to do to our regulations.

But maybe we also had this on, let's say, traditional databases. The data was sitting in locations that were traditional BA databases. We didn't really realize the impact that was gonna have on our cost, our licensing model for those databases.

Traditional databases don't scale the same as more modern data practices like a medallion architecture. So if you are running off a traditional databases, now your licensing costs go up, right? And this is another sort of challenge here.

And what do we talk, talk about? We call this a pear-shaped benefit. At first it gets great, then people are using this.

Then our portfolio managers come to us and say, wait a minute, our costs just went up, right? Be on our underlying infrastructure. These are, again, risks that we need to understand and manage.

Then again, regulation comes in. We don't have the provenance, we haven't followed those things. Now we've gotta step back.

Maybe we've gotta shut some of these features down, right? Because we don't, we can't have them. Regulation just came about that said, we need to have provenance for these tools.

Now we start reducing the benefit. We don't want this to happen. And I will share another thing here.

This shadow ai, before we had a term called shadow it where people would swipe their credit cards, go out and then figure out how to do things on the cloud. Well, that was fine. If you look a large organization, one of our customers has around 90,000 employees, right?

Um, and in that organization, they had about four or 5,000 people that were working with them to get things to the cloud, right? That's, uh, still a lot of people. But Microsoft just recently did a study and said that 70 to 80% of knowledge workers are utilizing ai.

So now imagine that you've got a 90,000 person organization and you've got 75 to 80,000 people utilizing AI to do their jobs. If you thought that shadow, it was a challenge at finops was a challenge, then shadow AI is going to be a significant challenge and something, a risk that we need to take into consideration as practitioners. Now, with that, I'm gonna go into this mergers and acquisitions use case.

Now, again, remember, this can apply to other projects, but mergers and acquisitions is something that regulated industries do. A lot of banks do this, right? So a lot of activity in the financial sector.

So I thought that this was a good primer to see how ai, the latest in AI can apply to mergers and acquisitions, and how can AI help you with that? So taking a look at the approach and some of the components, um, for mergers and acquisitions, we start with the strategy, right? Is this a strategic fit?

We conduct market analysis. We target, right? We do due diligence, we structure how these things might work.

We do integration planning. We look at regulation, right? And these considerations, what's gonna happen post, post mergers, contingency plans, and, and a lot of planning go into these beginning stages.

And I'm gonna step through how AI is gonna assist us in each one of these things. We then move into target screening, defining out how we're gonna screen the criteria for these acquisitions, conducting our research, right? Identifying target gaps, performing preliminary financial analysis, assessing strategic fit, right?

Evaluating synergies, right? Conducting risk assessments. Maybe you can start to think about while I go through these, how AI might assist us.

These are very human intensive tasks. There's a lot of resources that go into doing that. It takes a lot of time for our customers and the organizations that we support to do these activities.

How can AI uplift this significantly? There's a significant amount of these things that I just talked about that AI can assist us in now, and I've talked to people that are doing this. Many times what happens is, is they've already got an existing project workload and they get tapped on the shoulder to say, Hey, I need you to assist in this merger or acquisition.

It has come from the top board on down. We're doing this, and well, what happens to my existing project workload? Well, we unfortunately, in many cases, you've gotta do that too.

So we're not talking in this scenario about replacing people, we're talking about helping people and turning one person into 2, 3, 4 x more capable of doing things, leveraging tools like ai. Let me continue to step through these and then we'll move into how AI can specifically and pri precisely assist you. Obviously, we then need to look at due diligence, right?

Is this the right due diligence for it, for accounting, for legal, for regulatory, for compliance? Then once we go through the transaction, right? We have to go into very detailed, we have to make sure that we didn't miss any steps.

So that's gonna take people looking through the work that other people have done. This is a key capability that AI can assist us with. Now looking at the actual integration, how we're gonna bring all of these features and functionalities into our organization, AI can help us do inventories and things of that nature.

A lot of things AI can assist us with in a very large complicated projects like mergers and acquisitions. So now I'm gonna step through a methodology that you might look into that is specifically aligned towards mergers and acquisitions. But remember, this can be used for many other large scale applications as well.

What we, we talk about, we talk about an MVP, right? Not a most valuable player, so to speak, or a minimum viable product. This is really the most valuable product, right?

As we're building out our use cases, and these are workshops that we can assist you with or that you can do on your own. Um, but what do we wanna look at? So for this particular mergers and acquisitions, all of those tasks that I talked about, some of them are monotonous, but they're, some of them are very high, high value.

So we can look at productivity gains that AI can bring to us quality improvements. Again, AI can go back and check and check and check and check and check, right? Whereas if we have people doing this, they only have so much time to go back and check other people's, uh, work to make sure things that are validated, QA appro appropriately.

But AI can assist us in consistency and, and really help us with less human errors, cost savings, less human inve intervention, only having the person do what they need to do. Time savings, repetitive tasks, long running tasks, reporting improvements, less time spent on reporting, gaining more insights, automating visualizations and legal compliance. Like what I mentioned, if you're going a merger and acquisition in a specific area and the law changes, you might have to have a consultant that's assisting you on that.

Well, guess what? The AI can keep track of the legislation and it can highlight if you're doing any mergers and acquisitions in a specific location and can say, flag this particular project and say, these projects are good, but this project in this area, there's been some changes in compliance restrictions and legal restrictions for these acquisitions. You wanna know those things immediately as you're going through your due diligence.

AI can assist with these things. So once we do that, we really always try to start with the easy things, right? Identifying the AI automation quick wins, right?

Um, like scripts and templates and queries into databases optimized and adjusted. AI can collect insights from internal databases or internal knowledge basis. A lot of times when we're, we're doing a merger and acquisition, we're giving out documents to people and we're asking 'em to fill out these spreadsheets and things of that nature, they point to a document repository.

There may be hundreds of documents in there that somebody needs to go through, and AI can read all of those documents, synthesize what's in the documents, and put them into a format so that we know in this mess of unstructured data, we can pull that out and put it in a structured format and have the AI validate what's there, what's missing, what's not, and QA it. And another thing it can do is, as a document changes, somebody updates the document. It can't, it doesn't have to just tell you the document is there.

It can actually read through the document and tell you what was changed. And you as a practitioner can then see, ah, this is green. This person actually did do all of the things that they said in their milestones.

The AI assisted me in validating that. Then we look at things like ERP and, and, and databases, like SAP, like I, me mentioned quick wins and interacting with these databases, automation of environments, inventory environments, automated report generation. These are sort of the quick wins that AI can give you.

Now moving on up to things like NLP. This is also, by the way, in the AI guidebook from Prince two, um, looking at identifying these things like reading comprehension, writing and translation. A lot of these things in multinational corporations come to us in different languages.

AI can assist us in translating those into our native language, which is very, very important and time consuming from some, um, perspectives, right? Project specific chat bots, chat rooms and teams. Listeners, uh, all of you have been accustomed to pump someone, having a teams listener come in, but sometimes the transcripts are a little off.

It doesn't really know your jargon well. We can uplift the AI so that it really knows specifically precisely where we're at with the project. What's important for us?

Who can, we can assign action items to. We can have these sort of chat room listeners very, very good at our specific tasks, document management, right? We can also create mailbox ingestion agents that we can send an email in and say, Hey, look at this.

What do I need to do here? We can start dropping things into these mail agents and the AI can interact with us. AI can send emails on our behalf, and then when the email responses come back, it can synthesize that and help coordinate things with us so that we don't have to take a ton of time in meetings.

We can have AI running these things down for us. Again, chat bots and agents for one-on-one conversations. Many times when we're onboarding people, some of these are skills that we need to have.

AI can really know and become subject matter experts on what we're doing and can make the skills gap and the time it takes to onboard people much more streamlined. Identifying high skill, high experience, re things that are required sometimes, especially in mergers and acquisitions. We need to bring people in that need to be onboarded with very, very high skill sets.

Or we have people in the organization that have very high skill sets with people, new people coming in. Again, we have this onboarding, but with people in our organization, they're busy doing other things. Well, AI can become a subject matter expert and assisting us in managing these things so that our high skilled people aren't wasting time.

They're really focusing on the important tasks for the project, and the AI can understand these things and can coach the new people coming in to make the onboarding process much easier. Now we're moving into things like ontologies, which I will share at the end. Um, but again, these ontologies are more like, uh, allowing us to assess and double check things, understanding our gaps, understanding compliance.

We can also create these advisors, coaches, right? Frequency, a frequently asked questions for ai, making it very easy for these things, assessing our existing regulations, like I mentioned. And when new regulations come in, AI can understand and share with us the impacts, right?

And assist us with risk assessment and mitigation, which is a big part of our projects that we're looking at and how we manage them. So then again, a lot of times we're doing some low human added value tasks. They're just very, very cumbersome.

Formatting documents, mangling and joining data sources, updating KPIs. Some of this work is very, very easy for ai. And we don't have to have our people focusing on these monotonous tasks of formatting things appropriately or mangling things, uh, data sets.

The AI can look at these data sets, figure out how they join and make those efforts very streamlined and focused, and make it so that we're not doing redundant tasks for formatting and things of that nature and create visualizations for us. Now, with that, I think we're coming, as we start coming to a close, we wanna focus on some other things like QAing an end to end process. Again, like I mentioned, document management.

We have a SharePoint site or a series of documents. It could even be code, like I had mentioned before, code agents. And we can look through and find out all of the classes, all of the functions, all of the methods, everything that's in this repository that we may need to migrate.

That's something that the AI can do. Updating our KPIs for us and dashboards. A lot of times our practitioners spend a lot of time doing these updates.

AI can assist us with that. Again, now we're moving into something a little bit more complex, so I'll just touch on this, but as you go through a process like an m and a or a complex long tail project and ontology is what they call a first order logic. One of my friends, uh, Seb, who's, uh, our chief, our lead AI engineer, he always talks about Harry Potter as an example.

And he says, like, we have a rule that if you're a wizard or you're a witch, or if you're not, what are you, you're a muggle, right? That's first order logic. We can have the AI and teach the AI how to understand this first, first order logic and make sure that it's following these things, right?

Um, entities and relationship constraints, dependencies, goals and objectives. Are we hitting the goals? Are we actually hitting them?

AI can assist us with that. Defining rules and constraints that AI should follow up with or that we should follow up with or enforce. Now, with that, we then come down to sort of the, the very bottom line as we go through with organizations.

These are sort of two by twos where we can take all of the things that AI can do for us and the use cases that we come about up with, uh, for our organization. And then we put these in a matrix from time versus value, human cost versus skills and complexity versus the ongoing running cost of the tools that we build. Now, this is where you really will probably want to align with an AI practitioner for, for them to share with you the actual complexity.

But another thing that we talked about before was the ideation and concept sort of hackathon. So when you go through this as a primer, this should get your interest going on, what you want to accomplish, what could be quick wins, what can AI actually do for you? And then to lay that down on a series of two by twos like this to see where the cost and the value, the human cost and the skills, and then overlay that to the running cost, the total cost of ownership of what we put into place.

And again, this is something that you probably want, uh, a, a practitioner, an AI practitioner that knows the technology to overlay this for you. But these are things that large organizations, um, are going to be doing. You may have already been a part of these things, but if you haven't, these are some of the things people are asking.

Where should I start? This is a very tried to be a very comprehensive way for you to know where to start and how to contribute to the organizations that you're supporting. This is an an ontology.

Again, I know this is very complex, but I'll, but guess what? AI created this. I didn't have to create this AI ontology.

All I needed to do was give all of the elements of the m and a process to the ai, and then it figures out the relationships. Now we can validate this, but it has done a tremendous amount of work for us. Then I wanna leave you with this end-to-end knowledge agent scenario.

This is a very more complex, um, tool that we can maybe, uh, connect offline and talk about. But this one is for, um, global supply chain projects. Now, this was specifically built for, um, organizations that were trying to find out if subsidiaries in their supply chain were operating in sanctioned locations.

So what does it do? It searches and it finds all of the available information for all the constituent companies that we're using in our supply chain. One of these has hundreds and hundreds of companies.

Then they have sub-companies and sub-companies and sub-companies. Well, those sub-companies need to have warehouses where they're storing the stuff. Do I not have an address for that?

Where, where's the address of that warehouse? It has to have a, an address. Well, gimme the address.

I need to validate that address. Is this in a location that has sanctions on it? Maybe, maybe not.

Why is it missing? The AI can assist us with this. And up here at the front end, the person who's asking the question, we even have an AI that knows who you are and it can generate what you want it to generate visualizations or communication with you.

And this down here, we, as you can see here, this retrieval, augmented generation rag, along with on our ontology, becomes the subject matter expert to assist us with that. I think I want to just pause there and thank everybody for your time today. Again, my name is Joseph Enix.

Um, and I really appreciate you spending the time here with the skill up, um, for Prince two and, and understanding what AI can assist you with and what your customers may be asking you to do in the future with AI projects and how you can assist them getting started and to manage what AI can do for them as well. Thank you very much.