PegaWorld iNspire 2024: AI and Automation: The Digital Trust Builders in Omni-channel Customer Experience

In the age of digital transformation, creating a seamless, personalized, and contextual customer experience is not just a luxury, but a necessity. In this session, we will explore how AI and automation coupled with Pega's digital messaging capabilities helped a leading communication company achieve stunning cost savings and deflection rates. Throughout the presentation, you will hear from both the client and the ProCom partner about how AI decision-making capabilities enabled them to predict customer behavior patterns, build customer trust in self-service, and create integrated touchpoints across all channels that provided a consistent experience and offering.

Transcript:

- Look folks, we're thrilled that you have decided to join us. I appreciate it's been a long, hopefully inspiring, fulfilling few days here at PegaWorld. I really trust you've had a great time. I don't wanna say we've saved the best till last, because my other speakers would frankly have words with me, but I'm really glad you've chosen to take time in the schedule to come to this session, because genuinely, I think this is one of the sessions where you're gonna walk out of this with a real thought of, okay, I can use this practically, I could, you know, really consider using this in my operations and my organization. And that really is the goal of today's session. To introduce myself, my name's Simon Thorpe. I think a few of you might have seen me already, but I'm the, I lead our product marketing for customer service and sales automation businesses here at Pega. And we have a phenomenal session ahead of us, really digging into the importance of making digital and those digital experiences for our customers as frictionless and personalized as possible. And you can see behind me, I have a whole team of people who also think very passionately about this topic. So over here, I have Chris and Luke from Metronet, and they're gonna do a proper introduction in a minute, but hands up here, have people come across Metronet? Are people aware of the incredible growth of this business since 2005? One of the fastest growing ISPs out of Indiana. Very proud of the growth that they've done. And they've worked hand in glove with AJ Joy and Jason Salzman, who are from the ProCom consultancy business. And they've had some phenomenal success, leveraging both AI and Pega's Digital Messaging capabilities, some really practical use cases, which has had some phenomenal deflection rate and cost saving outcomes that I think we'd all be very, very interested in. So folks, we're gonna have time to quiz the guys at the end. We'd love it if you could save questions till the end. This is one of the highlight sessions, so we've chosen to record it and we're videoing right now. So if anyone's a bit camera shy and doesn't want to get on the microphone, I can completely understand, just bear that in mind. But we will have plenty of time for questions afterwards. I might throw in a couple of questions, just, you know, have a bit of fun, see how we get on. But without further ado, let's give the guys a big warm welcome, and I shall pass to AJ to introduce the ProCom business. So welcome guys.

- Thank you all for being here. And thank you Simon for helping us get to this day. I know there's a lot of planning and coordination, and along with, you know, Jesse laying on the Product Team for helping us on our journey with Metronet from the start, and partnering with us and allowing us to get here. So thank you for that, and hopefully we'll make it worthwhile. You made it to the last session, breakout session at PegaWorld, so congrats. Hopefully we're gonna be making this fruitful for the time that you're here. I know a lot of you have plans and things like that, so we appreciate you making it in and hopefully we're able to deliver on it. And there is a demo that we're gonna be doing towards the end that shows the implementation that we're talking about from an omnichannel perspective. So please do not miss that. We hope that you can stay through it as we get through the kind of the introduction here. Alright, so getting right into it. So ProCom, right? So we are a Amdocs company recently got acquired last year. We are now a parent company of Amdocs. Amdocs now are a global company. We currently are in a revenue of like 5 billion, a little bit closer to 5 billion in revenue, 31,000 employees across the board. So many solutions and clients, and it's helping us really drive that solution across globally. And we're proud of that and we're expanding our Pega reach. With prior to that, ProCom's been around 20 years plus prior to that. And we have made our brand being Second Surgery Pega Consulting Agency, which means we go in and then we look at things that were done bad and clients bring us in to make it right. Over the years we build relationship with executives at those clients that have now gone out and became Pega advocates that now are bringing us in from the ground up. And we have a couple of different clients that we have engagements with where the Pega application was brought in and bought the license. And then first off the ground, we're brought in to strategize everything from program management, road mapping, all the way to end-to-end delivery, as well as ensuring that we're delivering value as quickly as we can by making sure that we're covering all of our life cycles end to end from an interaction and Digital Messaging platform. So that's who we are. I'm not gonna go through every single thing, but we're in all verticals and anything that has to do with Pega, count us in, we'll be able to make sure that we're delivering that. We have domain expertise as well, along with everyone that are Pega certified. Alright, I forgot this. All right, so when we talk about ROI, right, this is something that we practice at ProCom just in general. So anytime we start projects, we wanna make sure that we're looking at our goals and objectives and then looking at how are we doing every quarter, even if the, at the end of the day, it's our responsibility to make sure that we're looking at the value and the ROI on what's getting prioritized, and provide strategy and recommendations so that way the client and us can partner on making the right decision. So if you look at this, this is an example of one of our clients that we've been with for over three years. And we look back, the savings that we see is 165 million, and that's huge. And then we have truck roll reduction, which is 70% truck roll reduction. So anyone who's in the telco business understands how much a truck roll cost, it's not in the double digits, it's in the triple digits. And when you look at how many truck rolls we can save by ensuring that we're alerting the customer on what's going on and making sure that, you know, truck roll is really needed by taking care, looking at the Digital Messaging platform and leveraging the logic and the orchestration. And sometimes it's just a manner of communicating properly. And so when we communicate to them and say, hey, we see that everything is working from our test and analytics, do you really need us to come out? And usually the answer is, hey, yeah, I forgot to call you guys. I didn't wanna wait on any hold, so no. And then we're reducing that truck roll, and there's instances like that that we're able to do at different levels. And we're doing all of that by leveraging the Digital Messaging platform. And we're doing that in a very quick manner when we start off the ground to make sure that we're doing the use cases on the orchestration engine, in a way where it can be leveraged across multiple channels. And so that's your SMS, IVR, and any kind of self-help platforms, like mobile, and so on and so forth and chat platforms, right? Going into the next slide here. So this is gonna talk about, you know, what were our objectives when we started out, right? So, this is our Metronet journey and this is what we're here to share about, which is what was our goal when we started out with Metronet in terms of what our objectives was? From an omnichannel, we needed to enable self-help at the IVR channel, chat channel, and web channels, but how do you do it in a meaningful manner where we can do it quickly and expose it into the multi-channel layer? And then how do you make sure that we're still providing great customer experience? So the, you know, a lot of times, we're so focused on saving cost, but then, you know, you're gonna jeopardize customer experience because you're talking about a customer base that isn't used to self-service, but now you're asking them to use self-service. So it's gotta work, it's gotta function and it has to be quick and to the point. And being able to think through that and figuring out how do you make sure that your experience isn't impacted but it's in a positive manner. But at the same time, how do you make sure that you're providing self-help services that isn't taxing on the operation side of things, right? And so that was one main goal driver of understanding what we needed to get done from an objective standpoint. And then last but not least, at the same token, Digital Messaging platform is one of the starting ground of expanding to a desktop channel, right? So things like web chat, all the chat queues, live chat, and all of those things where those things has to get taken care by a rep, and it's a live chat that has to happen. So what better place than to go from omnichannel approach of self-help service going into desktop where you're bringing in all of the interaction cycles that happened, so the agent knows exactly what messages went out, what messages came in, and then we're able to bring all of that contextual data into one desktop, and then the agent is able to help further without annoying the customer of asking, well, what message are you talking about? What interactions did you have in the IVR? All of those details are in the interaction manager, and we're exposing that in the unified desktop. So that way there's no repeated question causing further annoyance to a customer, right? So that's the desktop strategy there, and making sure that we have visibility across all channels and everybody is in one platform saying the same thing, using the same thing to communicate to the customer. Alright, without further ado, I wanna bring Chris on the stage, and these are our great partners that we've been able to work with. Both guys have been on the ground from start, from both of the projects that they're part of, and they've been more than great partners to us, and without them we wouldn't be on the stage. So come on Chris, this platform's yours.

- I'll pay you for that later. Good afternoon, my name is Chris Besner. As Simon mentioned earlier, I'm a Business Analyst for Metronet under the field operations side of it. I'm actually really excited to be here. This is my first time speaking at PegaWorld. So standing ovation's are welcome. Please by all means. Thank you. Yes, yes, thank you. Thank you, man. Just, I love this crowd. So who is Metronet? So we're actually one of the largest internet providers that are independently owned in the country right now. We've had phenomenal growth, and it's been just absolutely exciting. So right now we're in over 300 separate communities. We're in 17 states that our plant is in. I saw when Simon mentioned earlier, some people raised your hands, you know, if you're in one of these states and you didn't, come see me afterwards, you know? We'll see what we can do. Our family here at Metronet is over 2000 associates strong, and we're still growing. Now because we've had such amazing growth, it really kind of gave us the opportunity to take a step back and say, hey, what can we do to modernize how we want to communicate with our customers? What can we do to really give them the power to control what their install experience is like? So we knew we wanted to go down that path. The first step to doing this is we had to figure out how do we want to go about it. Up until that point, we always followed a waterfall approach. We would meet with the business, they would say, hey, we would love to do X. And we'd be like, okay, see you in three months. Come back later, deliver it to them, and hope that it was what they envisioned. Now since we kind of swapped to an agile framework, we really can work hand in hand with the business, truly ensure that as we are developing solutions for them, that it's exactly what they need and exactly what they want. Now this is gonna be super important later on as we get down here. After we've identified what it is that they want, we needed to make sure that they were trained and educated on the process. You know, this is gonna drive up excitement. You know, change can be scary sometimes. So if they're familiar with the product, if they're familiar with the solutions, they're going to embrace it more. After we got that done, we want to take a look at migration. You know, we have all these legacy applications that we've used for years, and we wanted to migrate to a more Pega-oriented solution, working with our ProCom partners here. Now in order for that to be successful, and this is probably one of the most crucial parts in any of the steps preparing for change, is to ensure that everybody is, A, excited for the migration, trained and educated on what that process is gonna be. And if you have all those factors in place, it's gonna lead into the very last point here, which is gonna be adoption. If your end users are excited, if they're absolutely thrilled about getting into the product, the adoption's gonna come naturally. If not, well then there's your opportunity to take a look, to see, well, hey, going back in the process, what is something we could have done better? And to talk about what actually went well and what some items that we could have done better on our end, I have Luke Wahl here, who will kind of cover some of those with you.

- Hello everyone. Thank you for being here today, and I would just like to start off with changing the slide. There we go. So as Chris said, my name's Luke Wahl. I am also a Business Analyst at Metronet and I've been with them for about three years, and with had the joy to work with Jason and AJ these last almost three years working on Pega projects. Now, my project at the company was DCCP, which is Digital Customer Communication Platform, which focuses for the most part right now on residential communication, and providing self-service opportunities for them instead of having to deal with a phone call with one of our customer representatives. This gives the company the ability to have those agents focus on more complex problems that require a human touch, while providing them more simple problems and tasks, provide those to our automation services, like rescheduling, cancellations, and actually doing some troubleshooting over the phone. So with that all being said, I'd like to talk to you, like Chris had said, on some of the experiences we've had with Pega implementation over these last three years or about three years. So with that being said, let's start with the first key point, which is growing pains. Now as Chris mentioned, training is a huge proponent when it's coming to moving from one program into a new program, this new program being Pega. And one of those is onboarding. Onboarding is a huge task for those who have experienced, you know what I'm talking about. For those who don't, or in the process of doing it, you could, it will happen. And with being said, onboarding has a huge task, which is moving everyone from your old system into your new system and making sure they have all the proper access points and everything they need to do their job. So I would say that's probably our biggest growing pain key point. I wanna keep it short. So moving on from there, we had what went well. What did go well? Now what Chris said mentioned earlier was moving into a more agile workspace. One of the things we did do well with that approach is talking, communicating with our stakeholders and keeping them in sync with our projects. By doing that, we were able to make sure we understood clearly what needed to be done and how do we do it. And if there's anything that needed to be changed, we made sure we made those appropriate changes in a timely manner. You also have to keep in mind some of our stakeholders don't have their projects done immediately. Sometimes it's three months out, maybe it's four months, five, six, so on and so forth. And making sure you keep that communication going with them is really a important aspect of keeping everything going and chugging and making sure you're making a good and well-oiled product for the company. With that being said, you also have to keep in mind velocity is an important factor, and don't wanna promise something you can't deliver. So making sure you have that velocity match what you're wanting to talk to your stakeholders about and making sure everything is roadmapped clearly is a very important aspect to that. So that's really what we're really proud of what went well. I could talk more about other things, but that's the one I really wanted to focus on, as I think everyone can either know about that or apply it to what they're doing right now. What didn't go so well was something I would say at least was something that we had a rough start on right when the project started for the first year, which is feedback. Now for those who may not know, well do know, feedback is important, but what I mean by feedback is post-release feedback. We would always take in feedback prior to a product being released and use that to develop what we needed to do and make the changes that we needed to apply, especially if it's something they didn't have in their previous product. Making sure we get that feedback from our stakeholders, and our agents, and the call center, whoever's front facing with our customers, that's important. Also our customers as well. But when we released a product we had to go to the next project, and sometimes we would get several requests to make modifications, make updates, or hey, this went well, can we add more to it? Hey this didn't go well, can we do a little change to make sure it's working as expected? We had a lot of that, and it came in through multiple different channels that were spread apart, spread around in our applications. So we had to make sure we centralized that a bit better and communicate with our stakeholders and making sure that all came to a consensus. So with that being said, what we did to improve that was making sure you get that feedback, get the most important stuff in the following sprint, if not the current sprint and anything else. Make sure you got the backlog and sure you don't forget about it. So that's some really important stuff when it comes to feedback. And something had a rough start, but we're pretty confident we're in a good spot right now after about two plus years, almost three years of doing this. So after all that being said, at the end of this, is lessons learned. When it came to lessons learned, like I said before, keeping in sync with your stakeholders, make sure they know what you're doing and make sure you're doing what they want. That's the most important factor. Also, providing clear communication with your devs on what needs to be done, and making sure they have the opportunity to come back to you with every single approach that they can be provided. And then making sure you communicate that clearly to your stakeholders so they know what's best for the company, and making sure you both are on the same page when it comes to that. So I would say that's the most important thing when it comes to lessons learned. And that kind of clears up the implementation experience so far being two plus years into the product. We're gonna move on to the next slide, which is some of the business benefits we've had with this. And I'd like to say a lot of these numbers, if not most of them come from the beginning of when we implement these products. The first one we have is customer device reboots as Tier 1 support, which is 71,000. Now this is something we were really happy we implemented last year, and this is our troubleshooting automation via phone. So for a lot of people when you call in and have an issue with your product, whether that be your internet or something else at your house, you usually have to go to a technical support team. That's been an established thing at every company, including our own. And one of the things we wanted to provide is for the simpler solutions, that we can automate and can remotely do something with, we wanted to create an automated flow. And with this was our customer device reboots. So it's our troubleshooting call in, we find your information, we helped check your system, see if it needs a reboot. If it doesn't, if it's a more complex problem, we'll take you to an agent. If it's a simplistic problem we know we can finish in service with a reboot or a rebuild, we'll go down that route instead. Instead of having you to get pulled into a call representative that, you know, may take 10 minutes, may take seven minutes. So this is something we've done that's helped increase our call center performance and increase call deflection. We're really happy about that. The next thing is something I had a brainchild of with Jason, which is our live chat system. So that right now is at 34,000 customer correspondence and growing every day. We're really proud of that. And so what that is for those who work in SMS, you got short code, you got long code. We use short code for this and just text agent to our short code, and you will be provided a menu and you talk with, you reply with the option you want, whether that be billing, or care, or technical support, and we'll get you with a representative via the same system web chat uses we use through live chat, and we're really happy about that. We've had, like I said, 34,000, so that's 34,000 people outside of calls with agents who can handle three to four customers at a time instead of one person on a phone. So I can't do the math right now, 'cause I'm not a human calculator, but if you did that divided by four, you can see the difference that is divided instead of 34,000 phone calls. So the next one I want to talk about real quick is 25,000 call deflected via outage notification. So this is something we're really proud of as well. We've been working a long time with our Network Automations Team. They've developed a way to detect outages. The two that we're really strongly supported right now is full pawn outages and any isolation outages. So with these, we were able to get Kafka events into our Pega system. They provide us the subscriber IDs. From there, we look up their phone numbers that they're supposed to get outage notifications through, and we send those appropriately to the ones that want it. And to the ones that don't want it, we won't send it. And we also put up phone intercepts as well. So this is a way for us to improve our call deflection rates during outages, which is something I guarantee you no call center likes to deal with. And the call center has been very happy to see this, and like the previous one, 25,000 and growing, as we do have outages during storms and weathers and tornadoes. So I'm really happy about that one. The last one I'm gonna cover is 17,000 appointment reschedules facilitated by via IVA and SMS. So when I first came to the company, the only way you could reschedule an appointment is call in, talk to an agent or go on your customer portal and try to reschedule it there. There was no other way of doing it. And one of the things we wanted to do is keep it within the flow of communication. And one of those things is we do have a point reminders we send via SMS. By giving those key words to them, they can reschedule right there and then where they have the reminder, and we've had a lot of success with it. It's affected a lot of calls that could have happened. And we've also recently implemented about a year ago, a little over a year ago in our IVA, which is our phone automation system through Five9. Pega is the brainchild of that, like it does with our troubleshooting automation, but we run it through IVA, and we tell the IVA what to do and all that jazz. And with this we've also been able to reschedule people's appointments through there instead of having to talk to an agent, which has been a real success. We're really happy about it. I will say, when it comes to phone automation, one thing we've hugely learned, if you haven't already dealt with this, is verbiage, verbiage, verbiage. Make it clear and concise and simple. The more simple you have it, the more straightforward it is, the easier it is, and the higher success rate you're gonna get. So back to one of our points, feedback is key. Listen to your feedback and you can easily improve your performance of your system. And with that being said, I have two more key points that we would like to discuss, but I would like to hand this over to Chris, as these are two from side of the side.

- Yeah, so these are actually two that I'm really, really excited and really, really passionate about. So starting off we were able to save over 9,000 repair deployments. Now the benefits of this is twofold. Not only are we completely reducing that maintenance cost to the business, which I mean, who doesn't like that? We're also opening up that capacity so we can use that same amount of time for either other customers who are in need, or to continue to grow a business by providing installs to customers who are interested in our services. So that's an absolute benefit to both our customers and to the business themselves. The last one is absolutely exciting. You probably never guess what it does by the title. We really wanted to facilitate communications between the customers and the technicians that are in the field. So we provided the way for them to send a message to the technician to let them know key information that's going on. For example, a customer may say, hey, I'm gonna be 10 minutes late. Little Timmy fell on the well again. Okay, at least now the technician knows he doesn't need to cancel that job. Timmy's just up to his normal antics, he can go ahead and get started. So by doing this, we're really able to kind of focus on that CX experience and really kind of drive that to the next level. But we're not done yet. We want to continue looking forward and see what else is out there, and what else we can really do to push Pega to its absolute limits. And for that, I have the one and only Jason Salzman.

- Thank you Chris. So we heard AJ give us some great numbers in terms of savings across the board. We heard Chris and Luke describe these numbers here that we're looking at, total value to the business. As ProCom, we actually, we have Metronet standing up telling us how great our Pega solution has been. So we really feel privileged to be in that position. You know, we've worked hard to get there, but I'd say, if you ask me some of the keys, how did we actually make that happen? There's two, I would say things that stand out. One is the solutions that we're describing here. The foundation has been Pega Digital Messaging, and we really believe the Pega team has done a fabulous job with Digital Messaging. It gives us the backbone for all of these things that Luke and Chris went through. So for us, we use Digital Messaging to provide consistent customer conversations across channel. So by a conversation I might be like, reschedule your appointment or help me pay my bill. Automated conversations, and we're doing that in the customer's channel of choice. So Luke mentioned a bit about IVR, a bit about SMS. One of the things that's really critical for us is that we build a process once, and then we use it in the different channels. So that's really the big win when we go into our designs and we say, hey, make sure we're following and we call it the x4, just to kind of keep us on track, make sure that we're thinking about all the channels, we're not just building it for one specific channel. And I would say like if you're, you know, using Pega, this is really one of the big discussions that we've had with, you know, from a design perspective, the IVR vendors will come in and say, oh no, we've got a great solution for building a conversation in the IVR. The SMS guys, oh, we've got a great way to do the SMS, and you know, using SMS. We say, look, it is a no brainer. You build it once in Digital Messaging, we plug it into all the different channels and the channels that are to come. And it's a really, it's an easy discussion, but we need partners like Chris and Luke that'll stand behind that decision and help us, you know, really believe in that vision. And that's gotten us to be able to do things that we thought were gonna be very difficult. Like Luke mentioned, the troubleshooting, we were able to do that, you know, very fairly quickly using the Digital Messaging. So I'd say we look at the successes that the guys are talking about. Number one, Digital Messaging. Number two, the fact that we're really thinking across channel when we do our implementations. So what has this gotten us is gonna, it's allowed us to save millions of dollars for our clients. And we've done this for multiple clients, with Metronet being the most recent. And for us, our goal is very simple. We want to provide the customer the best experience possible in the channel of their choice. So very simple, that's our goal. So we are very, we were satisfied with ourselves, AJ and I were sitting back, what are we gonna do this summer? So fortunately, we are not done. So we have a challenge. So our customers, they're making millions, great customer conversations, but that's not enough. So they challenged us, they said, guys, we think we can make our conversations better using GPT technology. So we've heard a lot about AI here at PegaWorld, so can we make it better? So, we did, we kind of had to take a little bit a step back 'cause we gotta ask yourself like, can GPT be accurate enough? You know, there's costs associated with GPT and what do those look like? Is it really reliable that we can use it for every customer conversation? And then a big one is security. You know, that sometimes people overlook, oh, don't pay attention. Oh, security is absolutely fundamental to our solution. So, we had to think about these things, and I just kind of pulled out some kind of recent headlines, you know, specific to ChatGPT. You know, one, ChatGPT recovers following an outage. So we need a reliable solution. We can't have outages, or ChatGPT spat out gibberish. We certainly can't be spinning gibberish to our customers. And if you actually go to the ChatGPT page, it'll tell you ChatGPT can make mistakes, consider checking important information. So, we've got to take that in context. So we've been talking about AI, we've been talking about all the great things about it this week, but these are the truths that we have to deal with when we want to use it in a production system to handle all of our customer communications. So I'll just kind of quickly show the, what our goal is. So we want to have a natural language conversation. The backbone for this natural language conversation, and there's all sorts of AI technologies. I have this discussion a lot. My personal belief is, for what we do, we are gonna have multi-step business processes. Digital Messaging is gonna provide that. GPT is gonna provide the natural language conversation. And as I said, our solution needs to be accurate, reliable, you have secure and we've gotta keep those costs low, 'cause if you look at like GPT4, those costs can go up quite quickly. So what does our goal look like? So I have a, this is a sample conversation, this is not the demo. I know AJ promised a demo. This is just the sample. So we have a customer, Emily, she's contacted us, we see she has an upcoming appointment. Nice and natural, how can I help you? So Emily says, "Can you help me change my appointment?" So it's in her language, we're not asking her to do anything specific, just hey, as if she was talking to a real agent. So we detect that the intent is rescheduling. So he said, "Great, how can we help you reschedule? "What date would you like, right?" And she said again, in her format, "Do you have a Saturday?" We take that Digital Messaging, we look up the available appointments, we provide it back to her and we say, hey, here's some appointment choices. Would you like one of these? Again in her language, "Can I do the 15th at 3:00 PM?" She selects it and we change it. And so that's our goal. We wanna have this nice natural conversation driven by Digital Messaging. So what does our solution look like? No surprise, Digital Messaging is the backbone. We use a specialized GPT API that we plug into Digital Messaging, and we kind of call our architect, just amongst ourselves, so we kind of know what we're talking about. We call it like a GPT-Oriented Architecture. What that means is that we're gonna use GPT to evaluate all of the customer messages. So any customer message, we're gonna pass that to GPT, GPT's gonna help us understand what does the customer want to do, and then we'll be able to continue on the process. And so, you know, the diagram here, I won't go into tremendous detail, just to kind of show we are using, yeah, it's all of your Pega assets, they already have, it's all your services that you already have. You're just using Digital Messaging to expose that. So with that.

- I'll add one more point to that, right? So when we had this challenge, you know, all the context that we learned in the keynotes with, you know, Infinity, 24 2, and all, we, you know, those things weren't available at the time, right? So we had to figure out how do we make the current framework work, but you know, the non-Infinity Pega applications and figure out how does it work both ways. But as we all learned, there's a lot more capabilities within the Pega platform that's already embedded. So as Don said, this is a fluid ever changing, but with Pega we can evolve, and this architecture is gonna be ever evolving and making it more and out of the box as possible.

- Yeah, I totally agree with you AJ on that. And I did have one person ask me, well what happens if Pega provides us in the product and you don't need your specialized API anymore? That's fantastic. We don't want to be in the business of specialized APIs. We want to be in the business of providing solutions to our customers. So with that, I'm gonna show the demo of Digital Messaging with GPT. Now just to kind of set it up, it's gonna be our same customer, Emily, and she's gonna start by rescheduling her appointment. So if you're able to see, so. So the first thing that we're gonna do is we're gonna just do some basic authentication. So nothing terribly exciting here. So we want to start the conversation. So we're gonna say, you know, "Can you please provide us your basic information?" And our customer's gonna be Emily. So she provides her information there. Now we say, right away we kind of detect, oh, you've got an appointment. So we give her some basic information. You have an appointment coming up. Now, very natural, "How can I help you today?" So we're putting it in her court, we're not asking her to give us any special words or anything like that. Now she asked us to reschedule, we take her message, pass that over to our specialized API, API determined her intent was to reschedule the appointment. And now we come back and we say, "Hey, of course we can help you reschedule. "What day would you like?" And now she's gonna reply in the language that she wants. "Can we do the first Monday in July?" So very natural, we use again, Digital Messaging to determine what are the options for her. So we say, which appointment would you like or would you like more different appointments? Would you like to keep it or just give us a different date? And so Emily reviews these, she wants to do July the 2nd, sometime in the afternoon. And then we say, "Fabulous, we've updated your appointment". So we used again, GPT. Everywhere you see messages, it's going to GPT to interpret. So we say, "Anything else that we can help you with?" Now she's gonna tell us, as Chris noticed, the tech note. So she's got some information for her technician. We're gonna pass that to GPT. GPT's going to tell us, hey, the intent of this is actually to create a note for the technician. So let her know we've done that. "Can we help you with anything else?" So now she's gonna switch context just a little bit. And now she wants some billing information. So again, GPT determines her intent is billing. We provide her some basic information about her billing and her payment. "Anything else we can assist you with?" And we talked about accuracy a little bit. We talked about reliability. And one of the big things is you have to give the GPT guardrails, so you can't let it go off the rails and start talking about things. I think that's one of the things a lot of people are concerned with. And so this case, you know, Emily's trying to trick us a little bit, you know, "Is there any show us to see in Vegas?" To try to see, okay, I'm working with the bot, you know, have these guys, you know, done the proper guardrails? And we have, so we limit what the GPT will actually respond to. And so in this case we let her know say, you know, thank you, and very politely thank you, but we're unable to assist you with that. You know, let's keep on topic. "Anything else we can help you with?" And you know, no, thanks for the help. Again, we've been conversational from the beginning to the end. So we were able to do the reschedule, the tech note, some billing information. She tried to trick us a little bit. It was all nice and natural, you know, from everything that we did. So this is kind of an, it's an enhancement of what's already great with the Digital Messaging and in terms of the customer conversations. And you look at the savings that we've been able to achieve. Now this, I think just takes it to the next level where you really are having that natural language and providing that great experience for the customer. And that's when you talk about, you know, giving the trust in your systems to your customers, this is the kind of experience that really, you know, gives them that, yeah, I have confidence in this system, I have trust in this system, and that they'll continue to adopt this style of bot. And when you look at a lot that are out there, they're difficult to use. And the first thing you wanna do is try to figure out how do I get out of the bot and talk to a real person? And so hopefully we've got a solution here that will be natural language and people will want to use.

- Yeah, and speaking of trust, right, so we have responsibility with the stakeholders, the customers, and then as an implementer, making sure that we have proper guardrails. So how do you make sure all of those things come together? It's, as you saw, as Jason was talking about it and showing you the demo, is by making sure that the experience is seamless, and we're not having to ask stuff like, hey, can you put that in a format, MDD, and all of those things, right, where the customer gets frustrated, I just told them next week, right? And so being aware of those things as implementation team is what drives that trust. And so when you deliver those solutions that are solid and that's well thought of, that's able to bring everything together, and then they're able to see, okay, great, now we can trust this to do more use cases. And that's kind of the story that we have at Metronet where we're getting stakeholders now to add a lot more to the roadmap than where we started at, just because the trust got build over time because they saw that we thought about all of the various, you know, nuances of conversations that would drive a customer crazy and would say, you know what, I'm never touching a self-service platform at this company again, I'm gonna call in every time, right? So it can go go that way, but it's our responsibility to make sure that we're thinking through the use cases and bringing it all in and then touch building the trust with the stakeholders and the customers, and our adoptions rates are off the charts because we make sure that we respect customers' boundaries as well, right, from contact preferences channel. So how do you not make it all spam? How do you make sure that it's relevant information? Another example would be like autopay, when your autopay is set up and your credit card is expired, like a customer would wanna know before their service gets suspended that, hey, the card that's on the file for the last two years is getting expired. So would you like to, hey, wanna let you know that it is expired and there is a grace period for you to renew that rate. And then that appreciation ties all back to how do you love the customer? Back to the T-Mobile keynote that we all heard is like, how do you make sure that they understand that we love them and the experience that we provide? And we hope that giving you the context and the insight allows you to drive those conversations with your stakeholders, with the Digital Messaging platform, and making sure that we're doing it the right way, so that way we're not making that trust go away from the stakeholders, right? So thank you for all for your attention today, and we appreciate your time and we will give it back to Simon for the rest of the events.

- Let's give these guys a big round of applause. Didn't they do an awesome job? And I noticed he snuck in and he beaming with pride. We have Jesse, our product leader, who develops all of this. This must be, you know, all your Christmas is seeing this come to life, Jesse. Awesome job, everyone. I love the practicality of this. You touched on it just there, AJ, that I, you know, when we were talking originally, and I'm a longstanding customer service contact center guy. When we were rolling things like Chat out and digital services, we didn't do a very good job. And we turned a lot of customers off, and it was because we weren't making it easy, we weren't making it personalized, we weren't making it natural. This is a prime example of how we build trust and adoption, and the continual cycle of, you know, using that on one particular use case here, but taking that same learning and the same way that that's been adapted into other parts of our business. That's the cycle that will move us ultimately the work away from bogging down our contact centers into getting customers to actually be empowered to self-solve. I'm super excited about this. We've got a few minutes, folks. Does anyone have a question in the room that they'd like to come up and ask? We've got some mics up here. Thank you sir. Come and tell us who you are and put your questions forward.

- [Dave] Dave Yuri with T-Mobile. Hey, congratulations for the outstanding work, and thanks for sharing results by the way. It's good to see the specifics. I was curious about network outage, and again, congratulations on seeing some lifts, and I should say reduction in calls. Would appreciate if you can elaborate a little bit more on the execution, right? Very time sensitive, wondering if you guys use CDH to determine, you know, who to message, and you know, anything you can talk about like with respect to channels, right? It was a notification, but was it SMS, email, digital?

- Right, so when it came to our network outages, we don't have CDH right now, but when it came to who we sent it to, so when we get those Kafka events, we'll have like, for example, like a node that's down or a terminal that's down. Network automation's job is to provide us the specific subscribers that are impacted by that. And then we take that and look up their phone numbers and we notify them based on their preferences. And also we control that by when we're allowed to send them, when we're not allowed to send them. So we send those out immediately when we know it's an outage and we are basing that off of other filters and factors to know if it's a legit outage or not. So not every, just because we get an event doesn't mean it's an actual outage. So we have filters in place to make sure we only send what's needed. Outside of SMS, we do, if I didn't mention this, but if you were to try to do SMS agent to reach out to us, we would also notify you there in a message saying, hey, you know, we're still more than happy to help you, but just so you know, you are in an outage, just like you would in a phone intercept. We have phone intercept queues for that as well. And then also on our customer portal website, we also have banners in place to notify them if they're in an outage. On top of that, we also have flags in place on the agent end. So if they get a call in, even before the customer even says, hey, my internet's down, ah! They'll see the flags and they say, hey, are you calling in about any network-related issues? We know that there's an outage impacting your area. And then that kind of drive the conversation going forward to try to reassure the customer that, you know, the agent understands what's going on.

- Yeah, one thing to add there, there's a full lifecycle involved with it too. So it's not just the initial notifications, it's the updates, it's the, hey it's up. And so there's notifications that are logical that makes sense in a meaningful manner until the outage is cleared. And then there's another use case around refreshing the network, because it may be just needing a router reboot of some sort. If they still say, hey, I'm still out, right? And so there's logic built in around those things too.

- Yes, and we do notify them when they're out of an outage too, so. We don't just tell 'em they are in one, so.

- [Dave] Thank you, appreciate the detail.

- Yeah, thank you.

- Alright, we've probably got time for one final quick one. I know we're getting to the end of the day, but does anyone want to jump up and ask? I think you've blown the socks off everyone. Blown the roof off. Guys, I think you've done a wonderful job. I really, really do. It's been an absolute pleasure working with you and taking this to life. You delivered it so beautifully. I know some of the guys are hopping on flights very quickly, but if you want to chat to them after the session I can put you in contact. There's plenty of ways that you can get in touch with Pega, and we'll facilitate that for you. Is anyone sticking around or are you all fine?

- Yeah, we'll be here a few minutes.

- Jason will be here. If anyone wants to come and have a one on one conversation, then by all means come up, but we'll leave it there. And once again, thank you all of you, Chris, Luke, AJ, Jason, superb job.

- Thank you.