Pular para o conteúdo principal

We'd prefer it if you saw us at our best.

Pega.com is not optimized for Internet Explorer. For the optimal experience, please use:

Close Deprecation Notice

PegaWorld | 36:42

PegaWorld 2026: AI in Action: Must-try Agentic Use Cases for Next-level Customer Service

Discover how agentic AI is elevating customer service. This session highlights practical use cases – like intelligent self-service and personalized support – that you can deploy quickly for real impact. Learn proven strategies, see real examples, and walk away with actionable insights to fast-track agentic AI in your organization. Perfect for anyone seeking next-level customer service innovation.

Good morning everybody. Welcome to the best post keynote session you're going to come to all day until lunch. In case you didn't know, you're here to learn about how we use AI in customer service and just the places you want to start. Those must use use cases and how you actually make this AI thing real in CS and what that really means going forward. So before we jump in, I do want to take a quick poll. How many in the room are on the business and service side of customer service? And then how many are on the technology side? It's actually a pretty good split. I'm impressed. 5050. Yeah, not too bad. So knowing that we have somebody who's more on the technology side and somebody who's more on the business side, I'm Brian Daly. I work as a senior manager on our business excellence team, which essentially means I take new products, work with early adopters, establish delivery patterns, and make sure we get wildly successful as these new products come into the real world, and not just in the demo world.

Yeah. So. Hi everyone. Andi Mutlow solutions consultant based out of EMEA. Um, in my role, I get to speak to loads of fantastic customers about all the great ideas that they want to do with our product, but I also get to build really cool and geeky demos and proof of concepts for them. Um, and as you can imagine, in this world of AI, I'm having a lot of fun with what we've got at our disposal. So one thing, and I know we get all the examples of AI all the time, we're used to it, but for fun, we did the headshots through AI. It's kind of ridiculous how small my hands, how how giant my head is and how great Andy looks. So I'm only taking it slightly personally, but as we talk about how we use AI in customer service, this is just such a high level example of the same exact prompt, same pictures, same formats, new thread.

So it has a little bit of history about me and what I do, but two drastically different outcomes, not necessarily what we want to be doing with customer service when we're trying to provide quality interactions and experiences for our customers, especially when they're during bad times and we're trying to get it better. Providing that consistency. And so how do we actually start leveraging AI to get away from the different chaos and introducing that risk of bad experiences? Do something a bit more in control. So we hear all the time about how AI is an amazing tool that's going to remove manual work and empower humans to be much more efficient at delivering extra value. An AI can be really, truly amazing, right? But how do we decide who does what and how do you use it? Because sometimes you might not use it, right? And because you were told to use AI.

It seemed like a good idea. You might not know how to. And that's why we're fired up for today, because we're going to talk about what use cases to start with. So step one is to recognize that AI is a tool, or it's many different tools that all come together, in fact. But it is important to make sure that you have the right tool for the right use cases. You also need to make sure that the person who will be using that tool is enabled on how to use it, and how to use it in the right way and the right, safe way. I have a two year old son who loves using his hammer in places he shouldn't. Never. Good results. Apart from having the right tool, you need to make sure you have the right thing to work on with that tool. In our case, it's data.

So as we specifically want to talk about this, as you're thinking about how you're providing customer service and you have this new tool, this isn't going to be the silver bullet that magically solves all your problems that already exist. You do still have to address. Are there data issues? Are there process issues being able to modernize, going and revisiting what is potentially wrong or what could use some optimization and then making sure you do have the right tool and the data is correct for that tool to work with. So how do you get from this chaos, which even in the best scenarios and best call centers and service centers I've seen, it's still chaos. Customers are hard. It's a hard job. So how do you get from that chaos to consistency? So for autonomy to be effective in customer service, it really needs to be grounded in three things.

Firstly, process. So that's what define what needs to happen in what order and what done actually looks like. When we talk about process here we don't mean a static flowchart. In practice this means the AI is operating within the constraints of the case, whether that is directly executing inside of one part or as a larger journey, or as one step on the way to an outcome. The case owns the state of the work. It knows what's already happened, what still needs to happen, and what done actually means. The AI isn't deciding that the case already has. That matters, right? In customer service, because customer service work is rarely just one step, a billing dispute and a dress change or somebody filing a complaint. These are the pieces of work that have to move through a sequence towards a resolution. So the agent isn't improvising its next move.

It's advancing work inside of a structured resolution model that already defines the correct path. It shouldn't sound new, right? Good process normally leads to good outcomes. We're providing a tool to empower that, enhance that, and enhance the experience for both your representatives and for the customers. Secondly, we've got policy here. We define what is permitted at this moment and what is not. What the system is authorized to do right now and under what conditions. Now, in most AI approaches I've seen recently, the rules sit on the outside, the agent does something, and then a separate layer is called to check whether it is acceptable, which is a really risky model in customer service. So the agent isn't choosing anything that theoretically could do at execution time. It knows what it can do, when it can do, and it's choosing from the options that are permitted right now and under the right conditions.

This is a fundamentally safer pattern for you where invalid actions are not even on the table to start with, they're excluded from it. And because these policies are explicit business rules, they're governed by the business. And thirdly context who is the customer? What's already been verified? What do we know about them? What's happened so far, what they've done previously over the last week, the last day, the last minute, and then how confident are we in what we know? What matters in service isn't just what the customer has just said. That context is persistent. It carries across terms, it carries across channels and across handoffs so the system can behave sensibly and the situations get more complex. It can pause, maybe because confidence isn't high in what it's doing. Maybe then we need to hand it across to a human to keep it in the loop, and it can escalate that and send across that required information for us.

And if the responsibility does move to the human, that person picks up on the same case with the same context and not a blank slate having to start from scratch. Without context, an agent might follow the process and obey policy and still make the wrong decision because simply it lacks the bigger picture. And so this is going to tie us back to where if you don't have these three things, you can still do stuff, right? Nobody in this room can say they have great process, great policy, and great content across the board. That's all interacting and working seamlessly together. That's what we're trying to get to. That's why we're all here. What AI does, apart from being a great tool, is interject this moment where you can have the conversations across the enterprise about the process, policy and content. This is the oh boy moment where you're like, what are we doing?

How are we doing it? And it's your opportunity to say, time out to do this. Let's revisit these pain points that have been existing and we've been ignoring or band aiding for three years and say, we can't go forward with AI until this doesn't work. And if you need help with that conversation, we're here to help. And that gets to where most organizations get stuck and why. And it really is around that, you know, the buckets of you could do, should do and will do. How many of us get stuck in that? Oh, we could do this. This could be a really cool idea. And then we try to move it to that should bucket and we should do it well. Now other teams get involved and things really slow down. And then trying to get to that final, oh, we will do it. And we're doing it is just miles and miles away because of the complexity that builds around that.

So what we want to do today and what you're all so excited for, I can see it in your faces. Let's talk about four specific use cases that you can leverage today that are quick, very great barriers or great levels to enter with and drive great value, great experience, but also provide insight and structure around getting to that golden state of good policy process and content. So we have our coach and agent. We have a quality agent. We're going to go through an analytics agent. We're also going to talk about a silo buster agent. These are the four you're ready for. I know you've just edge of your seat waiting for it. And with that, we're going to go to our newest member of the team. His name's Colin. So all of the agents you're about to see are ones that I've collected over time after speaking with those customers.

And they're all grounded on basically what we already have available now. OK Pega has got some great policies, some great process, some fantastic context. So it really is just a case of leveraging what you've got and pulling it together into these agents. And these are all solving real world problems that I'm hearing every day. Um, you're going to need to pay attention for on these because at the end we're going to get a bit interactive and we're going to see which one everybody likes the most. So this is Colin. He looks reasonably familiar I think. But what keeps calling up at night. Pressure onboarding new employees. He's got to do it quicker. He's got to do it faster. There is constant change to his documents, processes and applications, and he's often forced to act after the event when everything's just a little bit stale. So let's just have a look at what that means in action.

So our users logged into Pega and they've got a notification that's come through. So our coaching agent has discovered that there's something amiss and something needs attention. Now, in Pega terms, we could have sent this notification via Pulse. We could have sent it via email. We could have even sent using the new notification assignment engine that Don was talking about earlier. But what's happened here is the agent has discovered, based on the context that we have available, okay, via the chat transcript of the voice transcript that we have, that there was some dead time on the call and maybe we should have done a little better with our with our, with our conversation, with our customer. Now, Colin, he has all the access here available to what happened within that interaction. He's got our customer history, he's got that rich timeline of events, and he's got the summary that's been generated by the agent that then helps him understand what happened and what needs to go on next.

When we look here, we're going to see some of the suggestions that are going to be made to call in on how we can move this forward and what we need to do next, and what we'll see when it comes up is that the the the Coach agent has recommended that what we need to actually be able to do here is do some training. We can then integrate with our training providers and pull out courses that might be relevant. So how to master small talk one on one, something like that. But as with all agentic solutions, we've got to keep the human in the loop and we've got to allow them to try and make that final decision. And this example here, Colin decides, do you know what? Maybe that's not the right thing to do. Maybe I don't need to send my, my, my user off onto training.

And what we can actually do is we'll just send them a follow up. We can send across what the agent is recommended, how we can close those gaps and those small that small talk and then send it on and move on to the next scenario and. Thinking about this. And as you go, how do we take this to live? What? What does this level of effort look like? There's steps to it, right? We can show a lot. We're going to show a lot of functionality here. And for the next 30 minutes you don't have to do it all at once. Right? That's one of the great parts of Pega. But just technology in general is this agile approach specifically with AI. So these individual pieces, you might have more value in one with less risk, right? Just that even that little summary, you're not actually introducing risk.

If it is just a guided thing, you have it tuned appropriately where it is just providing gentle context so the rep can give a better experience. Those are great places to start. You don't need to be doing full self-service, which you can go see at the Innovation Hub. It is amazing. But we found over and over the business is excited about it. Technology is excited about it. But then that broader layer of risk compliance, how you actually do it the runway to get started using it is longer than we expected. The technology can move faster. It's the change management around it that takes more. So starting with something that's a little more direct, improves everybody's experience and doesn't introduce as much risk is a great place to start and why we started with Coach. Absolutely. So the next use case for our coaching agent is actually, how can we help our back office teams when they're actually live and working on a case?

So oh, sorry. My mistake. We haven't moved on the slide. Here we go. So somebody here has been assigned a case which is for their to enroll somebody in a loyalty program. Now they've got again all of the great information that helps them understand what happened in that case, the case history, the timeline of events. And they need to now go on and finish this process. But as we're going to see here, they're stuck. They don't know what to do. So they can reach out to the coaching agent. The agent isn't just looking at active things that it needs to be able to do based on events that have happened. It can work in real time alongside what's already happening. And we see here that our user reaches out and says, eek, I'm stuck. Now we spoke about context. The agent knows where we are within this process.

Within this case, it understands what needs to happen next. It can see the fields that we've got on the screen and the values that we need to put them. Now, of course, we can supplement the agent with extra context from Knowledge Buddy or any other training material. We've got to start to surface that insight to our to our users and help them move on. In this example here, our customer, because they're always super helpful. They said that actually, when it comes down to preferences on how to travel, they don't have any preferences. And we've got a problem here. We need to select one and our human agent doesn't know what to do. So they reach out, hey, I don't know what I put in here. And our agent says, hey, do you know what? Putting in rarely for this user is absolutely fine. Now there's still actually a little bit worried.

You know what if I put in the wrong value? AM I going to get in trouble? AM I going to break something? Should I escalate this to a manager? Well, guess what the coaching agent knows it's looked at our training materials and knows actually, do you know what? This is absolutely fine. You don't need to do that. Go ahead and off you go. Now we could take the coaching agent even further. We could give it the tools to actually automatically submit and complete that step for us. We could give it the tools to notify to our training team that actually, maybe we need to update our documents or any of those in between. Luckily, our users finished. They're happy. They've moved on to their next step, and we've had a real life coaching moment and meaningfully moved our progress on to the next stage. And as we look at this and we know the change management is a barrier, as we talked about enabling the people so they know how to use the tool correctly.

This to me is I feel like I'm on ChatGPT when I'm trying to figure out a recipe, right? You just slap it in and get something back. You're not you're not taking it word for word, but it gives you that guidance in your structure. Or if you do have a question about something you do know, it's that light piece of advice to keep you moving and keep you focused on what you're actually trying to do, driving the outcome that you're getting to. So when you talk about enabling people and getting better adoption of the technology, this feature we've seen, it's super comfortable because we're so used to it in our day to day lives at this point. Excellent. So the next person, Quentin again looks very familiar. Um, Quentin, the quality agent. Now, I'm going to be honest, as a father, I know I'm not allowed to have favorite children, but as a geek, I'm allowed to have favorite agents.

And this is absolutely my favorite one. Something that helps with quality, quality assurance. The problem that Quentin has is limited sampling. Customers are only really being able to assess maybe 2% of all interactions. The feedback is not representative of overall overall performance and high performing or struggling agents may be completely overlooked. There is an inherent problem with delayed feedback agents. Human agents don't get timely coaching. The opportunities for immediate improvement are missed, and customers may have already churned or are escalated. The final one is bias and inconsistency across a team of reviewers. Everybody's going to evaluate that interaction slightly differently. Different reviewers score different teams in different ways. Coaching quality varies across those teams, and humans may feel unfairly judged. So one of the reasons why I really like this quality agent is it's actually a bunch of agents. I love building agents. We've got an agent that looks at sentiment and understands, you know, was there a proper sentiment expressed one by our CSR but also by our customer as well?

We can look at empathy. Did our human agents show and demonstrate empathy towards the customer's situation? We can look at compliance. Incredibly important. We can feed our compliance agent with specific business rules. Business rules defined by your business to understand and make sure that we're doing the compliance right. And finally, intent. Have we done the thing that we promised for our customer? Is there anything we've missed here? How can we close that gap and move forward? Imagine. Right. Andy mentioned the 2% observability rate, but all the time it can be a little big brother, right? So you do have to make sure you enable and, and people understand what's happening in the background. Don't do this blindly. Agents get real upset when you just throw new metrics at them that they have no idea why it's happening, but you can't tell me, looking at those categories that you would love 24 over seven, just updates on how people are doing.

And while that can be a lot of data, having the agent actually manage that in, and you get to a point where you're really only throwing flags, flags for specific coaching moments. Or when you do have those specific initiatives for the month that you're trying to improve, you can tailor it very quickly with the agent to address that. It becomes an operational arm for you, for your managers, even for your agents to constantly get better with less effort to do so. So the quality agent is obviously grounded within a Pega case. Life cycle it has a structure that it needs to run through, so our agent knows what it needs to do, when it needs to do it, and in what order. So we ingest the initial interaction, we start to process it, and then we hand it over to our our sentiment agent for processing.

And then we go on to the empathy agent and we keep going like that. And each one of these points you can configure with business rules. What do you want to happen next? Maybe today you want to concentrate on compliance. Maybe tomorrow it's sentiment, but you've got the ability to configure the system how you want. And the great thing is, because it's a Pega case, you can then decide, well, once we've discovered a problem, what do we want to do? Do we escalate to a manager? Do we send an outreach to our customers? Do we flag it to every manager in the organization and get everybody on this problem to fix? But you've got that flexibility at your hands. And with new technologies, as we add new things, we like to make sure we stay safe. This is familiar, right? If you're using Pega for customer service or Pega in general, it's a workflow.

So while it is new and exciting and adding all this great capability, it's still rooted in the same thing. You don't have to do that much change management around it. You understand how it works. There is safety and security and the control of the workflow. And how that manifests itself and how you make these these decisions and these the judgments it's made. How you surface that to your users and your managers is perhaps even more important. So what we've done here is we've bubbled up the right information for us. We've shown the confidence that we have within this decision that's been made by the agent. We've got thresholds, whether it was a positive interaction or a negative interaction. And we've got, more importantly, the detail and the evidence of what we saw and where we saw it. So we're able to pass into the agents, the the chat transcripts, the voice transcripts, any other important piece of context, digest it, understand it, and then surface it up to the right people.

We can put this in a Pega screen. We could send it in an email. We could even pass it on to to another downstream system if we wanted to. That's amazing. Now, is everybody excited? I mean, this is pretty cool. You can imagine what you can do with it. I just I get fired up. So all of these agents have some additional instructions that we'd like them to be able to work with. The reason for that is because some of these things that we're looking at are actually quite esoteric. They're quite unique and specialized. And we also recommend and understand that you, our customers, want to be able to change this. Somebodies definition of compliance is going to be different to somebody else in another organization. You might want to be listening, looking for sentiment in a slightly different way with Pega and with the the rules that we've got available here, you can start to define that for you now.

I mean, that's as exciting as it is. It seems like a lot of data, right? Especially going going into the coaching agent coming into this quality. Agent AI does give us the ability and we've seen the AI slop, right? There's access to a lot, which means you also have to manage a lot. But what are we doing about that? Well, what we can do is we can put guardrails around all this. We can oh, excuse me, we can put guardrails around all of this. We can use that information to make sure that what we're using at the right time, as we said earlier, and against the policies that we've got in place to make sure that we don't go and use the wrong data in the wrong contexts. Okay. The next agent, the analytics HR agent, Alfred. Um. Alfred's pain point. He's got too much data.

OK can't find any insights. And maybe when he does find those insights, he can't always action them. He's drowning in this data. So how can we use an agent to help that? Now, we saw in that previous example an agent that had some quite good additional information. Well, hey, you can produce these without any additional instructions. So we've got a great set of out-of-the-box tools that are well described. They understand what needs to be done. So you don't need to provide the other info. But I had to have a little fun with this. Okay. I gave it all the tools it needed to get the job done, and I gave it three, three, well, four really simple instructions. I told it one that its name was Bob. I told it to. I wanted it to be quintessentially British in everything it does. And I said to it, wherever possible, use a bit of sarcasm just because that's kind of the really British trait.

And finally, I told it to be slightly passive aggressive. And you'll see the response that we've got here is it's come back to me and he's, you know, he's helpful, but he doesn't really want to be too much here. If you look here on the screen, the agent has described what tools it's got available. It's got tools to go and look at all of our service cases. It's got a tool to go and get all of the interactions. Now these are all out of the box. They are already available for you now to go and use. It's got tools to actually go and delve deeper into contact information. We could add more tools in here. We could give it a tool to send notifications to people to create cases for us. Again, these are all things that are available in our platform right now. It's just a case of giving them to the agent and then off you go.

And what you'll see here is that Alfred, typically he would look at a dashboard and he'd try and find maybe issues of, you know, ones that are breaching SLA, maybe cases that are close to breaching SLA. But what he actually wants to know is, is there anything that's flown under the radar? Anything that we've missed and we should have really been looking at. And what you'll see here is that comes back in a lovely passive aggressive way for me, but it discovers a case that's been languishing in our system for many, many days now. That's not ideal. We need to sort out how to resolve it so we can start to chat with, with the agent to understand a bit more about why this might have happened. What's the impact of this? What's our customer value? Can we make sure that we can rectify and turn around this system?

Now, where it gets really cool is when we look at our agents definitions, we can give it guardrails, we can tell to it things that it should do, things that it really shouldn't do. Now remember I told this agent to be pretty much quintessentially British and a bit sarcastic. Now we wouldn't want that polluting and spilling over into maybe a customer facing environment. So what we're going to ask here is actually, can we get Alfred, the the analytics agent to actually draft an email to apologize to this customer. Now, we can imagine all sorts of problems that we'd be in if it started sending emails that were pretty much as passive aggressive as this. But because we've got those guardrails in place, the agent knows when we need to be, make sure it's behaving properly. And what we'll see here is it will come up with a beautiful email that actually is entirely appropriate for the situation we're talking about.

And guess what? Yeah, we could have added a tool then to automatically send it on. We could actually add a tool to create a workflow to put it into one of our back office teams to review, approve, and then go off. And in fact, Alfred even offers offers here to take me through and automatically get the refund sorted for our customer here, which all I think you'll agree is absolutely fantastic experience. I don't know, Andy. I worked on the op side for a little bit. It only took me like three years to learn power BI, and then I'm pretty good at writing emails and trying to sort that out myself. I think we should probably just stick with the old way. Yeah. Who who has power? BI skills. And they love them so much. Yep. So see here. Okay, on to the next one. Um, back to my, my favorites and least favorites.

This is Max, the middleman. Um, there are silos between our front office and our back office. As much as we try to break them down, they're always going to exist to some extent. And that causes friction. There are misaligned goals and KPIs between the front and back of his teams. Communication gaps and volume volatility, where spikes in one area can have a massive impact on the other. Now, I don't have a fancy demo for this one. I've just got an image. I'm going to talk through what we could do. But this agent, what could it do? Well, maybe when our front office commits something to a customer, we can extract that promise that's been made. We can attach it to our case, and then we can make it as a tracked obligation. You know, the back office, they see the intent and not just a task.

When we see a surge in front office activity, we can predict workload using historical data and trends so that the back office gets suitable warning and not a backlog shock. The agent can spot stalled cases like that example we saw earlier, and then nudge or escalate with context. It can help prioritize based on the impact and or the customer value. So think back to that slide. Right. When we were getting from chaos to consistency and we're doing it with policy process and content. Now think of the agents we just went through and how much insight that can actually give you into those gaps of policy, process and content. So it is a self-fulfilling prophecy of as you get in and start using it, not only will it create the efficiencies help start solving those problems, but also help identify those problems more. The more clarity you have on those analytics, the more data you have constantly coming in about those interactions.

The ability to say front office and back office are now actually tied together in a much more efficient manner. You're getting to that point of, okay, we are getting more consistent with out introducing this huge amount of risk upfront without this year, year and a half long change management, which we should all be on that path, right? We are all working to constantly get better and transform and modernize and use all these tools. But getting back to that point of why do organizations get stuck? And it's because you don't have to boil the ocean. And that's why we're always trying to do these AI conversations constantly get up to, oh, this is a, it's a big thing. It's AI where you can take a step back and using things like this agent, right, that Coach, that quality agent, that analytics agent in our silo buster to say, we have a, we have a way to dip our toe in the water, but it's a big toe.

So we're going to get a lot of value out of it. We're going to increase our Csat. We're going to have less churn because our agents are happier. They feel more comfortable within a system. Just gentle empowerment of them. It's a thing that needs to constantly happen as we are transforming. We know everything's going to keep changing every two weeks. The world of technology is in a whole new level. None of us are used to. And to say that you have a grasp on it, you're lying to yourself and you're lying to everybody else. It moves that fast now. And so what you have to do is how do you make sure specifically in a service situation that the people are happy, the people are experiencing that technology in a good way. They do feel that empowerment. They feel the stress of their job ease away because they have the right tools and they are comfortable using them.

Familiarity with AI versus the black box. And that's not just transparency for regulation, but the, oh, I know I have a coaching agent that I can go to. I know there's an analytics agent that isn't there to punish me. Right. And this will get into the Change Manager about that, but making sure they know the technology is empowering them will then help them provide better experiences for your customers and the golden ticket on top of that, because we're now creating consistency through all this data. They're providing those experiences with less risk. And I don't know a better way to do customer service than that. So talking about moving at speed, um, I'm going to make a commitment now. So out of those four agents, remember I said I like building really cool stuff. Um, I'm going to make a commitment to have one of those agents available in our marketplace for you to download by the time that 26 goes GA.

So I'm going to have to move pretty quick. Um, but I don't know which one to build. I've got my favorite. So what we thought is you, our wonderful audience can pick that. Now we're going to go through one by one. We can put hands up, we can have some cheering, but. So let's start. Who wants to see the analytics agent. Put up on marketplace would use it right away. Okay. One okay. Um the coaching agent okay that's good. Oh lots of was that all the business people over there? Yeah. It was yeah. Okay. Um, the worst named agent in the world, the silo buster agent. Anyone? Oh, brilliant. Thank you. I might end up having to build two here if we get a die. And finally, the quality agent. Oh, it's between coaching and quality. Agent. Sounds like you're doing two. Sounds like I'm doing two.

Okay, so, um. Yeah, all of you here, please go and check our marketplace. When we go GA, you'll be able to download this, have a play with it, change it, learn from it, get some inspiration. Um, and I'd love to get your feedback after it when, when it goes. So final slide with that. Questions. Any questions from our audience. You have a brain beyond brains with Andy. So if you do have questions about the specific. Are you okay? Specific questions about these agents, but also the general concept of how to start adopting. We're open to all of that. If you could just come up to the mic so they have the audio on that. Awesome. Hello. Two questions. First one are the agents out of the box with a prompting so we can use it. And the second one is, can I create my own one with my own prompting and how it's going to work that out?

Yeah. So the first one, so these are not going to be out of the box. Okay. So I'm just going to provide them in the marketplace for you. They're kind of like an inspiration thing. Um, they will come with out of the box tools. So you're not going to have to go and create anything new. I'm only using things that are already in there. So anybody that's worked with Pega long enough, we've got a ton of data in there. We've got a ton of tools for doing things, and I'm basically just pulling those together to give you something that'll give you a bit of food for thought moving forward. As for you making your own agents? Absolutely. The one that we saw for the analytics, as I said, I did not give that any instructions. We've got a set of tools in there that are really well defined, well documented, and the agent understands what it has to do with them.

So when you give it a tool that says get service cases, all those service cases, it tells the tool will tell you what's in them, what to expect and how to use them. So you can go and create your own agents and just pull it all together with that. Okay, just one more and I'll leave it up. That's fine. So you can create your agents. It's coming from the marketplace. That's right. Data. Is it required to have access to the database of Pega? The cases in data externalized to S3 Dynamo. Is it compatible with the data pages MCP external services? Absolutely. Yeah. Data everywhere. Exactly. So exactly that. So yeah, we can use data that's in Pega. At the end of the day it uses a data page. And as you know, that data page can go off anywhere and get that. Our agent rules also have connectivity for MCP and A to a MCP coming in 26.

So you could start to bring in that full orchestration. So if you've got an agent that already does compliance, then just plug it in and off you go. Okay. Thank you. You're welcome. And as you start exploring your situations, please reach out to your account team, your solutions consultant, your CSC. We have experts who can help along this journey and talk about, right, what, what is built, what exists, what do you have to help build your own agents and go through that whole process? So please don't try and do this yourself. We have whole teams dedicated to it. Engage them and make sure you have the resources that you need as you go on this journey. Any more questions? I think everybody wants to go for lunch. Is everybody ready for lunch? Cool. Oh thank you. Thank you so much for your time, everybody, and enjoy the rest of your PegaWorld.

We'll be here if you have any other questions.

Recurso relacionado

Produto

Revolucionando o design de aplicativos

Otimize seus fluxos de trabalho com rapidez e com todo o poder do Pega Blueprint®. Defina sua visão e veja seu fluxo de trabalho ser criado instantaneamente.

Pronto para criar um Blueprint?
Escolha o mecanismo de reinvenção para suas necessidades.
Para fluxos de trabalho e design de aplicativos

Reimagine seus processos e transforme qualquer fluxo de trabalho em um aplicativo pronto para desenvolvimento, com confiança.

Pega Blueprint™
Para concepção de estratégias de marketing e CX

Visualize jornadas do cliente e estratégias de engajamento em todos os pontos de contato e coloque-as em prática.

Pega Customer Engagement Blueprint™
Compartilhar esta página Compartilhar no X Compartilhar no LinkedIn Copying...