PegaWorld iNspire 2024: GenAI™ Pega Cloud Exclusive: Paving the Way for an Autonomous Enterprise

AI won't replace people, but people who use AI will replace those who don't. I believe most of us are in this room today, because we feel that there is some truth to this statement, and none of us want to be left out. My name is Andre and I don't want to be left out. And I certainly don't want you to be left out either. This is recording from the latest Nvidia conference in Silicon Valley that happened this year. And this is Nvidia's CEO who kind of became this AI celebrity recently. Uh, looking at him. I don't think he actually enjoys this status, but he doesn't have any any choice these days. Uh, my, uh, my expertise is mostly about practical applications of generative AI and the cloud.

So I'm trying to attend all the conferences regarding generative AI that I can. Like the one like the one here. And by meeting all of these people on this conferences and attending all of these events events, I have learned two things about the generative AI. One Pega truly is the the state of the art, cutting edge gen AI provider. There is no doubt about it in my mind. I have spoken with a lot of our competitors and we are truly cutting edge. Two we are in the peak of the generative AI hype cycle, right? We all see it. We all want in, right?

We all we all want to do something with this. But very few people can actually deliver real value with generative AI today. Thanks. So According to recent Gartner studies, Generative AI is at peak of inflated expectations. And truly it is. We can hear about GenAI everywhere, at every tech conference, every tech company. Even today or yesterday on Kerim keynote, we heard GenAI GenAI GenAI right. But you will find people who are more skeptical to it. People who kept saying that this is just the hype.

There is nothing revolutionary in it. It's yet another technology maybe cool to play with, but nothing for industry, right? For real enterprise. Um, my name is Kamil. I'm responsible for Pega Cloud, uh, GenAI services. Architecture. And the truth is that exactly six years ago, I was among the same AI skeptical people. Uh, I shared exactly the same doubts when I was doing my masters in AI, and my master's thesis was about building an AI model for steel manufacturers to help them optimize the steel manufacturing process. And actually, that led me to things.

First thing was, how hard is it to train the LLM model? Because I spent so many months, so many computing resources and achieved not so brilliant results. The second thing that led me is that most most likely on my lifetime, I will never see generative AI working for enterprises. So I decided to switch to cloud. What can I say? After six years, life is a roller coaster. I'm back in AI. And I agree with Kamil. There is a hype for GenAI, but what Pega offers is real deal.

So let's quickly review our Pega Generative AI features, at least some of them. First up we have Pega GenAI Blueprint. It's like a private AI architect, almost like Kamil, that can craft the most efficient workflows for you, but not like humans. It does it in seconds. We also have Pega GenAI Autopilot that is designed to kind of hold your hand throughout the whole app development journey, giving you suggestions, making it easier and faster for you. And this one is exclusive on Pega Cloud free. But what if you are not on Pega Cloud free? Then the next feature is for you. This is designed to connect you to all the generative AI that we have in Pega Cloud free.

Next up we have Pega GenAI Knowledge Buddy. This one can instantly connect anyone from your organization with the information they need. And doing so, it kind of unlocks this collective knowledge of your organization. We also have Pega GenAI analyze, to analyze and summarize complex information for you and showcase only those vital pieces. Automate takes care of all the repetitive tasks for you so you can focus on the big picture instead of doing the same and the same again and again, right? Pega GenAI Coach. This is amazing feature that can help you navigate through complex workflows, giving you real time guidance and insights. And last but not least, we have Pega GenAI connect for all the other custom use cases you'd like to build on your own. And this is not all you heard about Socrates or CDH Assistant, right?

So we have pretty good suite of generative AI features. So my name is Yuka and I am managing the execution of Pega Cloud GenAI development. In my role, I mostly stay on top of this fast paced program to make sure we are moving fast enough to keep up with the crazy Generative AI world. And it's been a journey. And so let me just give you a quick story behind this picture. So this photo with Alan was taken during the Chairman Award winners trip this year that I was honored to be a part of. And honestly, I think that was partially possible because I use GenAI a lot in my role. And, you know, it helped me because it gave me this huge knowledge uplift and kind of evened out my odds for success while working with far more experienced colleagues. So I'm 25, and even though I'm Gen Z and kind of technology native, I've never seen technology like this before.

This is game changing. And I'm by this day amazed how it can boost my capabilities as an individual. And I believe that Pega GenAI can boost your capabilities as organization. And Alan put it best when we met. So we hear about GenAI all the time, and it's mostly more hype than a substance. But Pega Infinity is different. We offer this enterprise level generative AI capabilities that can help you work faster and smarter now. And so let's do a quick show of hands. Let's see how many of us by now used any of the generative AI features available on the market.

Show your hands if you did. Okay, so I'm not surprised that as many of us, but I think that it only shows how important this technology starts. It starts to be for all of us. So from oh, sorry, that's still me. And so today, with this excitement, we want to showcase to you amazing Pega Generative AI features and the benefits of Pega Cloud free and how these two combined gets you on this journey for the autonomous enterprise. So stay with us for this next 30 minutes to realize its value. So from all the GenAI use cases that we offer knowledge about, these are my personal favorite. To be honest. That's a real screenshot from my WebEx.

We use WebEx at Pegasystems to to do a daily communication, and this one is called People Body. This is something that our HR department built on top of of knowledge knowledge bodies. And I can ask it questions and it's giving me excellent, excellent answers as good as, as human would do. And by the way, it does not judge if my questions are stupid or if I make Grammarly mistake like in this example. Uh, so so here I asked about the internal process. Right. Our performance review, and it has given me the correct answer. Then I have to keep asking more questions about the amount of PTO that I can get in Poland. And yes, by the way, in European Union we do get a very generous PTO package.

Lucky us. But how did Knowledge Buddy knew these questions? The second one about the PTO. Sure, that's a publicly available information. GPT knows this stuff. But what about the first one? These are our internal processes. Well, our HR department have uploaded our HR how tos and HR procedures in a very secure manner to the Knowledge Buddy. And now I get to interact with it.

So you get access to best of these 2 to 2 worlds, right. You get to interact with this very smart GPT, which has access to wealth of generic knowledge, and you get to interact with your own internal knowledge. And Knowledge Buddy is intertwined. These two things to give you this great experience. And this is very simple use case, very simple use case. Yet it's very powerful. Just think about all low hanging fruits in your organization that you can solve with Knowledge Buddy. Think of all the hours that Knowledge Buddy can save you or your or your employees and Knowledge Buddy are available today. It's no, it's not part of this GenAI hype out there because it provides enterprise value today, right?

Just speak with your Account Executive we can spin up Knowledge Buddy for you in a matter of days, maybe even hours, even if you are not existing. Uh Pega Cloud client. Right. Uh, let's now look under the hood how Knowledge Buddy works to understand why it's Pega Cloud exclusive. So on this picture now, um, you will notice most likely Knowledge Buddy that are running on the newest and latest greatest Infinity 24.1, connecting to the GenAI services box on Pega Cloud. Each client benefits from its own isolated virtual private cloud. Um, so with that, with the VPC, you benefit from an isolated having an isolated and securely stored workload. But anyway, this VPC is running either on AWS or GCP cloud boundary. And however, via the GenAI services, it's connecting to the Azure hosted OpenAI models.

You don't need to worry about it at all because everything works under the Pega Cloud boundary. So it's given to you as an as a service offering. So you could probably keep hearing that. Pega GenAI is exclusively available on on Pega Cloud. To answer this question, why is that from the technical perspective? Let's maybe zoom in this GenAI services box. So GenAI services. Under the hood is an ecosystem of three major components. GenAI system, GenAI gateway, and GenAI Vector Store.

GenAI gateway serves as a link between your own isolated VPC and the LLM world. The gateway is primarily responsible for security, abstracting away all the infrastructure, complexity, and of course, making the operations easier. We've done a gateway. For example, we can switch very quickly between cloud vendors, LLM models, or do a minor version upgrades of those models. Right? Um, so what you what you see now, this is not yet part of our public offering, right? This is I just want to show you the number of experiments we are making under the hood on our internal R&D. Right. Uh, you can see that via the gateway, we are able to connect with llama three and Cloud three that are hosted on AWS bedrock, or for example, Google Gemini model that is running on GCP cloud vertex AI.

Um, I just want to show you that GenAI gateway is a very important piece for us from the architecture perspective, because it allows Pega to keep up with very fast changing LLM world without sacrificing any of the operations or security. Um, so another component GenAI vector store. Um, actually, this is a part of the Knowledge Buddy or or maybe a stomach when I think about it. But anyway, um, the first thing you will notice is it's connecting to a separate embeddings. DB. Right. So this is a separate database than the one you're Infinity is connecting to, even on the Pega Cloud. Um, and normally we don't show that number of details to our customers because they don't simply need to worry about it, right? This is as a service offering.

Everything is covered by you, by us for you on Pega Cloud. But it's PegaWorld, right? I think it's okay to show you some more meat. What is happening behind so Knowledge Buddy is actually consists of two very important functionalities. We call it knowledge ingestion and knowledge retrieval. So let's take a look how the first one looks. Um, let's say we have a document that was uploaded through the Knowledge Buddy. Um, let's say this is a document that describes Pega performance review process. Very, very long document.

I never read off. It's so boring. But anyway, you as a user, upload it through the Knowledge Buddy. What happens next? Then that document gets uploaded through the Knowledge Buddy to the GenAI vector store. Star GenAI vector star then splits this document into multiple smaller portions, which we refer also a chunk. Each chunk is then sent through the GenAI gateway service to the embeddings model. Open API Other tool, which is responsible for getting us back vector representation of this specific chunk. And of course the chunk is sent back to the GenAI vector store and we are saving it in the DB.

We are doing it, of course, for every of the chunk in the parallel for you, so you don't need to worry for that. So that is how you get your data in, right? You may ask me how you get your data out. That's actually very easy. Let's say you have a question. Let's assume that was Android question. Uh, when the performance review is happening in Pega, um, the question gets the GenAI vector store and the GenAI vector store Service uses its intelligence to find all the most relevant chunks for this question, and getting it back to the Knowledge Buddy. Knowledge Buddy then uses it as a context for the prompt that is sent to the GPT 3.5, a conversational model, and the question, of course, is rendered and sent back to the to the Knowledge Buddy. So I'm hopefully next time Andrei will not forget about my performance review.

That is, by the way, happening in December. Did you hear me? So that was just a glimpse of how GenAI services works for one client, right? But in Pega Cloud, everything is about the scale. Um, if you take a step back, we just noticed that was, you know, just a smaller portion of much bigger picture that shows the Pega Cloud scale. Yeah. And this is the scale that we are talking about in Pega Cloud, right? We operate across 30 regions based on two different providers. AWS and GCP.

So these large boxes represent regions. Of course, our clients are not that evenly distributed, right. But on the Pega Cloud, we have hundreds and hundreds of clients and hundreds of these virtual private clouds, these VPCs that Kamil talked about. And in every VPC there are tens and tens of Infinity instances. Right. So this is the scale that we are talking about here. And it all works seamlessly. We hear Pega Cloud folks are the are the are the only organization on the on this planet that runs Pega at this scale. That makes us experts.

Pega Cloud three is state of the art microservices architecture built on top of Kubernetes, and Kubernetes is the most modern cloud stack out there. Your workflows on Pega Cloud auto scales amazingly and restarts when they do happen. They take seconds Everything is designed in an auto self-healing way with micro updates are happening all the time in the background without you even knowing that they happen. The services that we use are already externalized, and you can choose between GCP and AWS. One of these two flavors and GenAI is available exclusively on Pega Cloud. So there are three ways for you to get access to this GenAI features that you that you've seen. One if you are, or if you want to become a Pega Cloud client. That's it right? You you can.

There is no technical limitation for you to get access to any of these GenAI features that you've seen. So that's the easiest path. That's the best path for you. That's the path I certainly recommend. But but if for any reason you must remain on prem or you must remain on your own cloud, then Remember Knowledge Buddy. Yes, they do run on Pega Cloud, but they integrate with everything that you have on prem or on your own cloud with Pegasystems or not necessarily. Pega because it also integrates with WebEx, as I have shown you. Now what about all other AI use cases like let's say GenAI Coach? Well, there is path for you as well here, right?

Prem bridge. Basically, my message for all of you here today is that there is path for all of you. And now I did mention GenAI Coach by no accident. I believe that's your favorite use case, right? Yeah, that is because with Coach, every complex process becomes instantly doable by anyone. And really. So let's think how it works under the hood. Under the hood, Coach integrates of course, Generative AI, but also reach Pega workflow context and your organization's best practices so that this real time guidance and insights that it provides to you is aligned with your vision and strategy. And so it really boosts productivity and make every employee your best employee as we see here.

So let's let's take a look at some examples. First up we have sales coach. And in this example as a salesperson you can chat with this coach. You know conversational way. And in this particular one you want to know how well are you following your selling methodology or what steps in this methodology needs the most attention to and so on. And very quickly, right away, coach gives you this data driven, personalized answer that you can leverage to act on it right away. No need to dig through, you know, thousands of documents or status updates or whatever it is, and then analyze it for hours. You can just act on it based on the the answer it gives you in the next example. Very, very simple one.

We have home loan summary Coach. And here you want to know how you are doing with this particular home loan. So what happened by now? What events was taken, what what was accomplished, what status changes happened. And again you get this step by step summary very quickly. And in this particular example, let's think how useful it can be when something went wrong. We need to figure out what. And actually it can help us trace back to this process very quickly and see, okay, in this step something failed or was successful. So really boosting again our productivity.

And based on those two examples, very simple ones, I think we can all agree that this is not a fair game to compete with someone who use AI tools while we don't our brains can't match the speed and precision of AI. But this is this amazing era when we can leverage this technology to support us like never before. And in Pega Cloud three, we have GenAI to support you as an organization like never before. So, Kamille, do you want to talk about support in Pega Cloud three? Yeah, some again, boring stuff. But, uh, let me show you the operations aspect of LLM and how it's, uh, done at the at the bigger scale. So instead of me talking to you, how hard is it or how easy it is? Uh, let's make an exercise. Uh, let's assume you are not on Pega Cloud.

If you are on Pega Cloud, you can essentially skip all this, all this picture. It's not applied to you. It's as a service. But you are not on Pega Cloud. You are running Pega, right? And you are Pega engineer. I hope we all do. And and you've been tasked with integrating your on premises application with GenAI. You won't have your own GenAI.

So the first question you need to have is actually pick an LLM model. So which model would you choose? Most likely Open API it becomes a golden standard. Now it's always on the on the on the top of it. Um, if it's open AI then you need to pick a model, right? The model version, which is GPT 3.5 or GPT four. You want to start small in this exercise. We don't want to get complexity of GPT four. Um, so I understand that.

Let's pick with GPT 3.5. Uh, you need to pick a flavor of that, which is 3.5 turbo or Turbo 16 K. 16 K might sound a little bit old, so keeping things simple. 3.5 turbo. Let's stick to it. At the time of development, all the last thing you need to choose is the minor version of this model, which is in this case zero 613 zero 613. Um, the code name is that it was deployed on 13th June of 2020. Uh, 23. Um, you go to your operations team, you ask them to deploy it for you.

Uh, you get the URL, you put it to Pega, build your own custom GenAI use cases. Everyone is happy. Six months later, you receive a deprecation note. 11 zero six is the newest better and 06306 13 is deprecated. You must upgrade. So what are you going to do? You go to your operations team, ask to click an upgrade button, right? Because it should be that easy. Um, they get back to you saying, hey, you know what?

You've been deployed initially in East US region in Azure, but now you need to switch the region because 1106 is only in west US going all this compliance certification. Talking with security guys, they are the hardest. So you need to justify this business justification. But let's assume after I don't know weeks you are done with that. Uh time flies. Six months later another deprecation note zero 125 from 25th of January. Now is the latest and greatest. You must upgrade. So again you go to your operations team.

And again, another surprise zero 125 is not available either in East US or West US. You must switch region to North central US. Do you feel this operation pain now? Of course you don't if you are on Pega Cloud, but if you are doing everything on your own, that's the hardest thing. And I didn't even cover all these other families like GPT four, 32, K four, turbo four and by the way, if you at any point decide to switch to GPT four turbo because you know, GPT 3.5 turbo, GPT four turbo sounds similar, right? Another surprise is US two regions that only closes one available for you. So there are other models which we use in Pega like image generation embeddings model. Everything becomes very complicated unless you are on Pega Cloud. Why?

Because in Pega Cloud everything is about the scale. This slide is actually outdated because it says 29 global regions. But the truth is, since yesterday, we operate on 30 global regions having more than 15 compliance certification running at the same time. 30,000 plus compute instances, most likely since the beginning of this presentation, We upgraded two environments of our clients and with zero downtime, of course, they didn't even notice our security. Our architecture implements layer Security with best in class encryption. All of our deployments are multi zonal, highly available with needed data redundancy. We run mission critical application of our clients for more than ten years, and our expertise spans for both AWS and GCP. And this expertise is actually something that guarantees your success with Pega GenAI on Pega Cloud. AI won't replace people, but people who use AI will replace who don't.

I would like to paraphrase this one a little bit, because I believe it's also true for enterprises. Enterprises that leverage AI workloads extensively just just have a huge competitive advantage over those who don't. Organizations who are on the path to autonomous enterprise are just better off. The Autonomous Enterprise is a self-optimizing business that applies AI and automation to decisioning operations and servicing across the organization. This is the age of autonomous systems. We see it everywhere. We see it in the cars where basically every newly produced cars have some degree of autonomy. We see it in shops where there are automated cashiers and we see it in enterprises. Of course, of course.

Let's get one thing straight. Running a corporation. Running an enterprise is way more difficult, way significantly harder than driving a car. So we don't see the same level of autonomy in the enterprise that we see in car. Sure. But it's here. This is the 30 years of enterprise evolution in one slide. And if you look to the left, the internet. Just imagine running an enterprise today without the internet.

I will risk the statement it's like nearly impossible. Why? Because if all other enterprises in the world are leveraging internet and you don't, you just can't compete with that. The internet is such a huge competitive advantage. And now generative AI and autonomous enterprise in the cloud are predicted to be as significant as and as big as the invention of the internet itself. So you simply cannot miss this train. And we here in Pega Cloud can help you get on this path right here, right now. And since we are almost done here with this session, let's just briefly recap it together. So you have all those and not only those great AI features available for you today.

Right now. It's not a hype, it's reality. And we touched today on Knowledge Buddy, Andre's favorite that can instantly connect you with the knowledge you need. What it does, it actually completely eliminates the need for manual search in your organization. At the same time, improving self-service and ensuring that everyone is up to date information. My favorite Pega GenAI Coach. That helps you navigate through those complex workflows, giving you real time guidance and insights that are aligned with your organization's vision. And let's not forget about Pega GenAI Kerim bridge. That will be your way to get the generative AI features that we have on Pega Cloud free.

If you're not there already, and with all those features, really it's exciting to see how we can boost our productivity. But actually, that leads me to another important point that I would love you to remember after this session. So Pega Cloud three takes care of the complexity for you. Again, auto scaling so it adjusts to the usage peaks. We have blazing fast restarts, so if services has any issues as they occasionally do, we Pega Cloud free restart them. So downtime is minimized. We also have self-healing infrastructure and independent micro updates. So instead of having massive system updates, we do those. We apply changes to each service without disrupting the entire ecosystem.

In Pega Cloud three, we also have externalized third party services for better security, stability and performance. And also you can make your choice based on your preference if you want to go with AWS or GCP. And with all those really great benefits, taking care of this complexity for you. I think still there is this one very cool benefit of Pega Cloud three so GenAI available there in its full potential. So thank you everyone for staying awake during this presentation. We are slowly heading to the end of the PegaWorld so I appreciate it. I appreciate your time. It was a great pleasure for us to be here with you. If you haven't checked out Blueprint yet, which is very unlikely, but if you haven't, just go there and and and play with it, it's really amazing.

And I will tell you a little secret. Under the hood it runs on Pega Cloud as well. So this was just a very high level overview of, of of how GenAI works. There's so much more to know which this is just what we could we could cover. So I'm super excited to hear what questions you might have, if any. So we could go a little bit deeper. So that would be time for your questions if there are any. Yeah. We have a question.

Yeah. Yeah. Um, so question about regional failover I understand. Zonal where you have the ability to go between regions. One of the reasons. You want to take this one, I think you are actually working on multi-region disaster recovery feature for Pega. So the question was about regional disaster. Yes. That's now on our roadmap.

Uh, that's coming in this year. Uh, so that will be also a part of our as a service offering. So even when the region goes down and you want to ensure you work out, very mission critical applications are still continue to work, but in different regions, that's also available in Pega Cloud. Yeah. Stats published on it as well. So everything is on part of our public roadmap now. Uh, this feature as of now is under development. I taking care of that from the architecture perspective. This one Cloud architecture, not the AI, so don't worry.

We have everything published on our public documentation, so if you google Pega Cloud roadmap, you will find a roadmap till the end of this year and you for sure find the middle related. And that's for multi-region disaster recovery. Multi-region. Right. So this is this is uh, this is an extremely rare thing, right? So basically AWS regions or GCP regions are are already designed in the disaster recovery way. There are multiple data centers under the hood. So disasters like that very rarely happens. But we understand that some of our clients wishes to have another level of uh, be prepared for the worst.

Does that answer your question? Awesome. Any more questions? Yeah. Please do. Oh. Um, so you mentioned Pega Cloud 33, and you mentioned Prem Ridge for integrating with private Cloud or on premise. What about Pega Cloud version? So we are actually undergoing a move, uh, update of our Pega Cloud to clients to Pega Cloud, uh, free.

I'm not sure how far we are with that. Probably halfway through. Right. So. So if you are existing Pega Cloud two client probably in the matter of the next months, I'm not sure where you are on on our schedule. You will be on Cloud. You will be on cloud free. So that's the best path, right? Just, you know, we'll just we'll just move you to cloud free.

If you are interested in GenAI, uh, we can just, you know, tell me. Tell me the name of your system, and we'll try to put you in the front of the queue for the for the next update for cloud free. Cool. There was a question there, I think. Um. My question was about, uh, you know, the this session, you said GenAI and Cloud. Right. So any relation between GenAI and Cloud particularly, are there two different topics that you want to discuss today? Because I want to know These GenAI tools are a part of the product as such, right?

But is there a necessity that we have to be a Pega Cloud customer as well in order to access any of those features? So, so there is a link between Pega Cloud and GenAI. So, uh, I know it may be a little bit confusing, but I will tell you just two statements. First of all, Pega GenAI features are exclusively available on Pega Cloud only, but at the same time you can get access to GenAI features if you are on prem, right? So there are right not compatible statements. So here's the trick. Right? If you must remain or must remain on prem, what what we what we do for you, we build for you a Pega Cloud system, right? So you become a Pega Cloud client in a in a small form where you get this bridge, for example, or Knowledge Buddy, and then you connect your on prem systems to this Prem bridge which is on Cloud.

So yes, GenAI is still exclusive to Pega Cloud, but we give you access through Pega Cloud to GenAI through your on prem from your on prem systems. And just to mention, this integration is as easy as getting the URL and adding it to your on premise system, unless this on premise is also on 24.1. Almost trivial, I would even say. Yeah. I got another question too. So I know, um, so I know how to ask this question, but usually. Right. You know the trend with upgrade, right? So we have an application on Pega Cloud right now that's on 86.

We are trying to skip Infinity 23, but upgrade directly to 24 because it is available. Right. But the usual upgrade pattern that we have seen is most of the business demands. We will have to end up customizing something and the customizations break with every upgrade. Right now that GenAI is come to that capability of trying to detect and understand and give relevant answers. Is Pega envisioning anything using GenAI right to predict and automatically fix things that might make migration journeys easier. Like, you know, like Knowledge Buddy or like Coach or anything else. Is there going to be an upgrade buddy or something that's going to help, you know, upgrades become easier? You know, so upgrades.

But I don't think we've thought about this one. That sounds like a very good idea. That sounds like a terrific idea. But what we thought about but not commitment, it's just internal people in Pega talking, right? So we have an entire effort of of modernization of the clients, right. So we have clients who are on prem on pretty ancient versions of Pega. Right. By by all standards. And how how do they how do they modernize?

So there are some discussions, how we could use GenAI to help with this with this process like maybe analyze app, maybe give some tips, maybe even do some basic things for you. So they're still there still is a necessity in the cases like that. For some heavy lifting for a client to rewrite some parts of the apps. But we are we are looking into the into the possibility. But nothing, nothing on the roadmap yet. But but yeah. Upgrade buddy. Excellent. Excellent idea.

Last question that I got. So any dependency on Cloud three for any of the GenAI features or even Cloud two can do it. You want to take that? Sorry, I didn't catch that. The Cloud three is there dependency on Cloud three for any of these features? Or if I'm on Cloud two and upgrade to Infinity 24, I still can use all the features. You must be on cloud. Yes, you need to upgrade to. Pega Cloud on Cloud three.

We can give you kind of like a playground to test if you'd like to test if you'd like to, if it's worth it to move for Pega Cloud three, and kind of use all those Generative AI features today. Uh, but yeah, most of them are exclusive to Pega Cloud three Knowledge Buddy, so you can get them without being on Pega Cloud three, as Andre mentioned before, and primaries. But this is to connect you. So this doesn't really count. Yeah. So other than. That to be on cloud three for getting these right. Yes yes yes. But there are, there are like I don't want to use this word like but hybrid modes like so Knowledge Buddy if you are waiting in a queue to be moved from Pega Cloud to Pega cloud three, and you are so excited that you can't wait another day, right?

We can build an environment for you on Pega Cloud three with Knowledge Buddy, and you can integrate them with your existing cloud two systems, right? So there are like a there are there are possibilities like that. I am in fact actually using this in the reverse. Basically we are on cloud two already migrate to cloud three. But the journey was not easy. We saw many, many challenges. Like many interim fixes, production was breaking and we are kind of working back and forth. I even have a meeting tomorrow, you know, in the evening to talk about what things broke in production because of cloud three upgrade. But we are on eight six.

We are not Infinity yet. Maybe I want to use this excuse or use this, you know, argument as an excuse to my business to explain why we moved to cloud three. Maybe, you know I can use GenAI only if I'm on cloud three, right? So. So first of all, I'm sorry to hear that. That's the first time I hear that. So? So I hope that's a good thing that that's the first time. Right.

But but yeah. Yeah. GenAI is absolutely great. Business justification or business excuse. Right. To convince your business to to move to cloud three and to and to a newer version of uh, of uh, of Infinity as well. Right? Right, right. So, I mean, the interesting thing is, the reason why we had to move to cloud three is a different thing that was not discussed here at all.

Something called immutability. That's a that's a security requirement from our enterprise. And Pega was not immutable before. For us to have immutability enabled for us on AWS, it had to be on cloud three. And like GenAI is just a blessing in disguise, you know. I'm sure we'll sort it out. So. So yeah, let's sort it out. Thank you.

Any more questions? Anyone? We have a question. Yeah. Question. So. So do you want to answer that vector DB who's our vendor? Well, actually, uh, that's a secret sauce at the moment. Uh, at the moment we are using the open source technologies.

Um, and uh, we keep it, I think, aside, because the reason we have a separate service behind it is actually to be to have this ability and agility to switch it whenever we see there are better performance, uh, test performance results on different, different vendors. Yeah. But but under the hood, deeper below the service that we have created is Postgres database with pkg vector. There are many reasons for that. Uh, first and foremost, I'll be honest with you. We have tons of Postgres expertise in Pega Cloud. So we went with something that we can provide the best value, operational value to you when we're making this decision to go with with pkg vectors with Postgres for vector data store. According to Google, there were better options. But today when you Google, what's the best?

What's the best database to store your vectors? PKG, vector and Postgres always end up in the top three, right? So we have made a good choice and we are very happy. But as Kamil said, we have built a service around this so we can quickly switch to something better when or if it emerges. But currently it is RDS on AWS or GCP with Postgres and pkg vector under the hood. And Knowledge Buddy. Is very automated. We have a. Yeah.

Yes. Yeah. So there are two there are tons of ways to to to ingest or upload your knowledge to the, to the Knowledge Buddy. But I will just divide them into two buckets. One you use knowledge management application. So you upload your documents. Everything is nicely tracked very very professional, very very enterprise friendly, I would say. And then there is another way, which is also enterprise grade. But it's but it's a little bit more hacky in a sense.

It's more like for sysadmins to do you get the ingestion API. So you, you basically can create your own scripts or your own tooling, right, to ingest whatever data you like. You decide on the size of the chunks. We basically I'm saying hacky because it's more developer friendly and sysadmins friendly because you have you have much better control. So you can ingest data regularly from, I don't know, SharePoint or your other systems. And we also offer a couple of integrations like that out of the box. And if I recall correctly, we even have some CLI builds around it. So that integrates directly with the Knowledge Buddy API Sanjay mentioned. Someone put it on top of that.

Yeah, yeah. There is option to clear documents as well because that actually happens. Right? You ingested something that you didn't want to ingest, right? So. So yeah, there are there are options to track it. There are options to remove it. To clean it. Uh, yes.

Absolutely. Uh, I'm not sure, but I'm pretty sure on, on our documentation page there will be a lot of details, but but I can, I can I can give you a name of the product manager who owns Knowledge Buddy. He's actually somewhere around here. So, so so he would love to have this discussion. Actually, I'm pretty sure he's in in, in in Innovation Hub in the Knowledge Buddy booth. Yeah. And with that I think that's, that's the end of the, of the session. Uh, we will be in the innovation hub for the next two hours. So if you have any questions, please find us there.

Uh, again, it was great to to have you here and, uh, yeah. Enjoy the remaining part of the PegaWorld. Thank you, thank you.