Amdocs Business Assurance Services

Make it amazing. It go. Make it diamond. Make it amazing. Thank him clap. Make him jolly. Make it amazing. Okay. Hi, everybody. Thank you for joining us for this workshop. We're going to be talking about how 5G is changing our world. First, we're to have Karl Whitelock from IDC talk about how the industry is actually looking at 5Gs, what is changing, what are the changes, how is this changing our market and technology. And then we're going to have a panel, us from Amdocs, talking about how to take whatever 5Gs is doing and actually take actions in business assurance. At the end, we'll try to answer all of your questions. So let's start with Karl. Hi. Today, we're hosting Carl Whitelock from IDC, and we would talk to Carl about how 5G is actually changing the telecom industry, and more specifically, how is this affecting the business assurance part within the the telecom industry. So hi, Carl. How are you today? Fine. Thank you very much. Thank you for joining us. I wanted to ask you, Carl, what what exactly is the effect that you see that 5G is gonna be having on the telecom industry? Well, I I don't know how much time we have to get into this, but, you know, we could talk for hours on this subject. I think one of the most important things to remember is this that 5G is changing the way operators do business. That's first and foremost. It's not just about changing the way technology operates or how the network works or how, you know, the technology interfaces with customers. It really is changing the way organizations do business because the network operator in the regards to 5G is going to be a piece of an end to end solution to more complex business challenges than has ever been in the past where it was all just connectivity related. So, you know, with that, there's going to be a lot of not just change, but I think complexity added to the overall environment as well. So one of the things we're hearing is the fact that, you know, in and then we had a lot of technological changes and seems that 5G is gonna be much more complex and bigger, is the fact that we'll have a lot of third party or or in in in agreements within telecom operators? Yeah. I I mean, I mean, a classic case is is that, you know, we've heard of something called multi access edge computing or MEC for short. And really what that is is it's bringing the enterprise edge as close to the to the telecom network or the telecom cloud as some refer to it as possible to take advantage of the latency aspects that 5G, you know, brings to light. So what that means is is that, oh, you could take like a, you know, there's some real good use cases where it's gone beyond proof of concept now. It's actual reality where a large, you know, hyperscaler cloud provider working with let's say, you know, a large, you know, 5Gs operator, they bring that 5G connectivity with the associated functions that the edge can deliver, whether that's, from a compute, and storage capability, or even some general AI that goes along with it. And then you have behind those two, you have a third party app developer that is associated with creating solutions for individual industries. And what ends up happening is is that the now the network operator is a piece of this solution, and therefore, if the operator is going to be the one that is taking the overall responsibility for the services that are being produced and delivered across the network, then partner ecosystems become a bigger and bigger issue. And that introduces some new challenges because this is an area that really hasn't been, can I use the word, really focused by the operators in the past? Yet, I think, as things grow, in the 5G world, partner ecosystems are gonna become a bigger and bigger and bigger deal. Okay. When looking at, you know, these really major changes that are happening, especially on the business side, what do you see as risks coming with 5G? We all know new technologies have new risks, or it takes time till they mature. But I think above the technology here, there are other risks that are involved. Yeah. Well, when you talk about risk and we and we kind of break it down in the traditional fraud management and revenue assurance aspects of risk management from the past. It was always you know, our our fraudsters trying to gain capability without paying for something and we had a whole series. The TM Forum went through a whole series of, you know, fraud use cases in that way. And on a revenue assurance side, we were always monitoring and measuring that the information about the use of the network was flowing from the switches into the billing system, and we were always making sure that all those data flows were were matching up properly. I mean, that's a great first start. That's what we've been in, and now we have to deal with this interaction with partners. And if I just stay on that subject for a minute, it's about onboarding a partner. It's about off boarding a partner. It's about understanding how much of the resources a partner provides that are actually used on an individual service, and it's going to vary service by service. And then how do we collect the insight that we need to know how to provide the right level of remuneration back to that partner? Because if we, you know, if we don't get it right from a monetization aspect, you know, these partnerships are going to go away very quickly. So it's complexity upon Complexity from what we're used to in that regard if that makes a if that makes sense and And I guess in that regard risk management means There's a whole lot There's a whole lot more places where Something can go wrong either by innocent interactions Something is forgotten something is overlooked something, you know, changes and doesn't get recognized as a change, or it could actually be from the result of nefarious actions by both internal and and external sources. Now we've been talking about 5G for a long time now, and if it's been a couple of years that we're talking, and it's always seemed to be something, you know, in the future, we'll be there. But how do you see it? We think that 5G is here now. It's it's something that you need to be ready for it now. How do you see Yeah. I I you know, I'm here in North America, and we've got a couple of carriers here that are, you know, knocking each other out. Actually, three of them that are knocking each other out trying to say, we have the best five g network. Well in reality, it's just still in its infancy from a consumer perspective. Is it another g from a consumer perspective in many ways? Yes. It is faster. It is better data carrying capacity. All the things that a customer you know a consumer would expect to to see, but that's where the difference or a similarity between five g and previous g's You know cease because five g now with its distributed aspects brings in that enterprise capability and that creates new solutions that are definitely important to business that ultimately gets shared down to the consumers, but you know at the end consumers are just looking for ways in which they can make their lives simpler and easier, but it's the business aspect of everything here that really becomes more significant so In some ways yes five gs is is going and it's operational. I think we hear some terminology out there. We've got You know five gs standalone and five g The current version of five gs and pieces is is not standalone and The difference basically is in the standards process and in the technology that's being used So it's that standalone environment is where things really get quite interesting, and I think even the most aggressive operators are in the early stages of implementing a standalone five gs network So I don't know if I'm answering your question, but it is and it isn't. Yeah. Yeah. And and how does real time connect to that? Did you see any aspects of of real time more today in five g and ever than than before? Oh, yeah. I mean, it's real time, and I think another term that we're hearing a lot about is something called dynamic change. And so the together, those two terms are are really implying that look, if if I'm going to meet latency standards around a defined service, then I've got to be able to monitor all of the pathways that are delivering that service from a network perspective, as well as from a partner interaction perspective. And when I see there's something falling away, I need to be able to turn up a new piece of the network in a software controlled manner, and take down the old piece in the same manner all without the customer being aware of what's going on in the background. If I can do that with the network, then I'm I'm truly operating in not just a real time sense, but a real time with with high value that I because I'm meeting those service level definitions that, you know, enterprise customers are now signing up for two. So that gets down to that quality of experience from the customer's perspective, which is, you know, a very, very big deal, especially when you're talking about use cases that are in certain industries where latency needs to be, you know, extremely extremely low, as in the case of maybe, you know, medical procedure being I'm not thinking about the remote surgeries, we'll get there someday, but I'm talking more about diagnostics in the in the health care field, being able to see information in remote locations as well as at the present location at the same time, and then maybe being able to make judgment calls from there. It could also be in sports where you've got a fast paced and like basketball or baseball or any of those sports where, you know, it's real time action all the time. We've seen it. We've seen that here in North America, especially in when it comes to the current football season. So you know, it's out there and you know real time is not something to be taken lightly because it it's now real time from making sure that the service is delivering what is designed to deliver, but also real time from the sense that we are collecting the necessary information to bill the customer for the usage of the, you know, the right elements that are defining that service. And that's not always an easy thing to do. It's not just call detail records coming off of a switch anymore. It's much more, much more. So this this this, of course, will be affecting any business assurance group that needs to now prepare and start working and building a strategy and a solution to to actually control five g, both on the revenue assurance of fraud or billing assurance because some of the the billing would be maybe not collected in real time, but building in in real time or invoiced in real time. How do you see business assurance group needs, or what do they need now in order to be ready when the full five g's would be out there and it's not so far away? I I think that business assurance teams need to be aware of some of these interactions with the network. I mean if we go back to my example earlier of the of the MEC based services, the multi access, what that when when we talk about multi access, we're seeing that we're going to have You know a touch of the network and a touch of let's say again hyperscaler or other partner delivery capabilities in several places that need to be then aggregated together to deliver ongoing real time quality experience acceptable services and Therefore I need to have business solutions that are also Monitoring and tracking in each of those places. So it's a it's a distributed net operations environment as much as it is a distributed network environment. Operations needs to come back together in some form or fashion to an aggregation point that would be then the control point back to keep everything in sync in sync. That is not an easy task. The opportunity of data becoming misaligned is tremendous. And so I think, you know, the the assurance teams need to be thinking about some of these new scenarios as much as the technology teams are thinking about how they're gonna technically pull this all off. And do you also see one of the things that that we're looking at is the fact, you know, we moved from billing per minute or different types of billing to more of a package. And now with five g, it seems that we might be moving back to some of the things actually by billing being billed or invoiced by event and not by package. Do you see that happening? And this will change, of course, billing assurance and other Yeah. You know, I think there's gonna be a long term evolution here. I think we're going to see some services operated by by business results. You know, have I have I met the SLA requirements? Is the experience acceptable? Did the business results? Were they achievable based on what what the conditions were defined to be? I think we've got this other other factor that's playing into it that as we cloudify the network, and I'm talking about, you know, the VNF and CNF NFV type world, we have to wonder we have to make sure that we keep in track those, those cloudified aspects because, we may see new billing, scenarios where it's not based necessarily on on, volume or on speed. It could be based on interactions with the cloud, you know, input output, if you want to call it something. That that that could be a new parameter. It could be a parameter based on meeting latency issues. So, you know, everything tied to a continuous monitoring of SLA defined experience. You know, am I within the limits of the SLA? And if not, what's my remediation? It's not just simply returning somebody's money in some regards. Certainly, it's going to have to be, at least not billing for, something when you're not meeting certain thresholds, but what else does it mean? So a lot of factors have to be considered here. If we want to summarize, what would be your recommendation for for a business assurance manager, you know, who's used to dealing usually with monitor consumers, some business aspect, but what would be your recommendation for someone to prepare for that? Okay, I guess I'd boil it down to a couple of things remember five gs is complex five gs also has a lot of moving parts one particular work team cannot possibly understand all of the interactions that are happening. I think you need to work with, you know, perhaps, you know, a partner of some variety that has some significant level of expertise in this area. Someone that you feel like you can trust to say, what about this? And what about this? Because there's gonna be lots of questions that need to have answers, you know, provided to them. Don't try doing this all yourself. You it's you're just not gonna get to where you need to be. It's an aggregation, and also don't think that what's in, the problems that are coming up in your organization are the exact same problems in someone else's organization. I know we've seen a lot of that with fraud management. Certainly fraud management is going to continue to behave that way, so nothing changes in the traditional sense. But it's the new stuff because the the the new solutions or services that are offered inside your organization are not gonna be the same as something offered in someone else's organization. So pay attention, but still have that interactive communication, not only with trusted partners, but but again, those that you've grown to trust in other organizations. This is definitely a team effort here. It's complex, and it's only going to get more complex with time. So don't try to do this alone. Well, Carl, thank you very much for this. It's it's been very interesting. Five g is really you know, it's it's it's great to live in exciting times, and it's really I believe also is really gonna change the industry for telecom. It's gonna it's in a few years from now, it's gonna be completely different than what we see today. And so thank you very much for your time, and see you soon. Okay. Great. Thanks for having me. Take care. Thanks, Carl, for sharing with us your views and thoughts about how five g is changing our market and both in technology and in business side. With me is Doctor. Gady Solotorevsky. We are both from Amdoc's Business Assurance Group. You also know us as CVIDIA. And we would like to talk with you about how to take these things to ground level. What do revenue business assurance sorry, managers and also analysts need to do in order to be ready for these changes in our technology and market, five gs, IoT, cloud and everything, how can we be prepared for that. Also, I would like to remind you that the chatbot is open, and you can ask your questions in the chat question box, and we will try to answer as many of them as we can as we go along. Hello. Hi, Hezi. Hi, Gadi. Gadi, perhaps we can start by some recommendations that you can give Business Assurance managers, what they should do. Okay. So I would say first is knowledge. Most of the things we are talking about are new, new to everyone in our business. It's not just new to revenue assurance, fraud managers or business assurance managers as they are. So first is to gain knowledge. We need to really understand what five gs is doing. Karl spoke about some of the things that is changing, but there are many more, not just on the technology side, how to get information, where to get information, how it is moving to real time, streaming data, etcetera. But there's a lot of business changes which differ between operator to operator, and we really need to understand exactly what is happening in order to be able to manage it and control it. So we need to know what the changes are. We need to do a very, very thorough gap analysis to really understand what we are doing today, what do we need to do tomorrow and then realize what is the gap and what do we need to change both on the technology, maybe changing our platforms, enhancing our platform, but also on the business rules we need to implement. And we need to build a methodology. The change is so big that it's not starting from new, but it's a new project. And we need to build new methodology of how we are doing our controls because a lot of things are changing and define a good plan. And we recommend to select a partner. No one can do it alone. No one can start now and say, listen, I will take my people, start learning. It would take too long. It's not the right way. You need to select a good partner with knowledge in the areas that you are doing and work with them to define the right strategy and the right methodology. And last but not least, we have to move. We were preventive. We were detecting at the beginning and then we moved to preventive. We need to move to predictive. We need to move to predictive technology, use machine learning AI. We need to move to real time because a lot of the changes that are happening now actually are driving us to do so. And Gadi, I talked now about machine learning explain a bit more what is the role of AI or machine learning in Business Assurance? Sure. But let's start with what changed. And as you can see in the TM Forum Business Assurance survey that is being released these days, the volumes of leakage increased drastically by over forty percent than two years ago. So something changed. What's this something? First, are seeing huge changes in technology, five gs, marketplace, cloud. Then we are seeing also huge changes in the way that customers behave. Everything due to COVID, a huge acceleration in the movement to self-service, doing everything digital. All this changed quite a bit the way that things work. So the technologies that worked in the past, rule based system rely mainly on experts are not fast enough, are not good enough. With this huge accelerated pace of change, we need something else. And this something else is machine learning and AI. It brings several benefits to the table. First, machine learning permits you to look at the unknown unknowns. You're changing technology, your customer change their behavior, you don't know what are your new risks. By using, by applying machine learning, can detect unknown unknowns. And this is a huge benefit. It really can be the difference between detecting a problem now and waiting six months until your detect it. Then machine learning permits you to generalize. You want to detect problems, but once you detect a problem, want to detect several things that behave in a severe way in order to detect also them. By applying techniques of machine learning like supervised learning or semi supervised learning, you can do it. But you also want to be able to predict and prevent the issues. And here also machine learning can help you. So the combination of these three factors, detection of unknowns, prediction, and the capability to generalize are huge advantage of machine learning and these are the things that machine learning brings to the table, not instead, but on top of the existing technologies to be able to first fight leakages. And second, and perhaps we will speak about it more later, to ensure your customer satisfaction. Jesse, with that, I spoke at high level, perhaps you can take it to the ground. Can you give us some specific examples? Yes, sure. So first, I would like to indicate that we took a little different approach when entering machine learning. We started our journey a couple of years ago, a little more than a couple of years ago. And we understood that we cannot come to operators and tell them, listen, we are a machine learning hub or we are expert at machine learning. Just give us the problem and we'll try to solve it because it doesn't work. It's very sometimes it's very hard to define the exact problem you need to solve and what exactly you are expecting from the machine learning models and and technology. So we took a different approach, we tried to understand, work with them to actually define specific use cases that we know what exactly is the problem, what do we need to solve and how can machine learning model can actually help there. Because one of the issues we found out that you run machine learning models, you get results, it's not always that clear what to do with the results, how to implement them in real life and actually get value out of them. So we took several we have now a bank of many machine learning models. I will talk about a few examples of what we're For example, we went with several operators and we looked at the problem that they have with bad debt. We have to understand that when we started our journey, it was not very long before COVID started. And COVID changed a lot, a lot of the behavior of customers, a lot of the changes within the operator. Of the things that we found out that there are more bad debtors than before. And it's not that there are more just more bad debtors. Different types of customers become bad debtors. Customers that you would never think would become a bad debtor in the past are now becoming bad debtors. And how can we now reduce the problem or how can we reduce the time it takes to actually handle those bad debtors? So what we did, we define a model that actually look at bad debtors in the past, supervised model, and try to profile what is the type of customers that has a very high propensity or likelihood of becoming a bad debtor and then mark all the new customers or existing customers what is their risk of becoming bad debtor. Now we need to understand, okay, so I put a risk on each customer. Now what do I do with this risk? This customer is now paying. It's not that he's not he hasn't paid. Or if it's a new customer, there's a risk. Even in ninety five percent, there's five percent that he will pay. So this is one of the examples of what do you do with the result of a model. So we helped the operators and what they did is listen, for one example, they took and said, okay, this customer has very high propensity of becoming a bad debtor. That means that if he doesn't pay his bill, the chances that I will recover the money are very, very low. You know, in a regular process when someone doesn't pay his bill, you give him another fourteen days and you contact him and at the end he owes you three months more or less because there is, you know, a chance of retrieving the money and also getting the customer to be back with us. But those customers, their chances are so low that cut them when they stop paying. So you reduce the exposure of bad, you reduce the bad debt, you reduce the time it takes. You don't have to go outside to sometimes lawyer companies and others to actually collect the money because there will be no money there. And we have an example from a customer that indicated using this model, they actually reduced the bad debt by almost half, by forty eight percent and reduced the time it took them to handle a customer from a hundred and twenty days to ten days. So these are examples that actually show huge value if you know how to run the model and you know what to do with the results, which is also something. And and that's for bad debt. We also did that for several customers to predict the propensity or likelihood of a customer to commit fraud. So you know better what to do with them. You can change the policy for these types of customers, how much you are willing to give them, how many phones, would you like a deposit and things like that. You change the process you do with this customer. It doesn't mean that you don't sell them. And the even bigger benefit that they get since you can mark those customers which are at high risk and usually with a model you can do it way more accurately than using the rule based, the number of rejections that they have went down. So they were able to approve more transactions that they rejected before because they had some basic rule based, you know, rules to actually decide who to sell to and who not. And the amount of money you're losing on the lost customer is much higher than sometimes the fraud that a customer might do. So at the end, this is the focus to actually increase the number of transaction you can actually do and not just pinpoint who are the fraudsters. And there's also we did models about abusing discounts, customers abusing discounts. Customers abusing discounts. We found cases that sometimes, you know, somebody doesn't pay a bill and you connect them and you tell them, listen, okay, so pay sixty percent of the bill this time and continue to be a customer, and then you learn that those customers understand the way it works. So each month, they don't pay and get a forty percent discount on their bill. So loyalty, we found problems in models about loyalty points. A lot of customers are giving you credit bonus that you can actually use. It's like real money. And we saw a lot of abuse with that. Lately, we did some models also on billing. When you're moving to billing, some of our operators are moving to what we call real time billing. It's not real time collection, but it's real time billing. So the customer can actually see very short after he did any type of change or transaction. He can see how this will be in the next bill if you go to the operator's website or to the app. And usually, you do bill QA or billing assurance when the cycle ends. So what happens if it happens a day after and you can see it a day after, so you need to change the way you're doing billing assurance to be more real time and actually use machine learning to pinpoint where exactly you think the problems will be so you can be more focused. So these are some of the examples that we're doing. Faizy, this really sounds great and very interesting, but two questions. How much it takes to implement a model like this? Is it years of work? Is it days? And the second question, I before I mentioned that one of the things that lead to machine learning is that everything is changing all the time. So you build this great model, but now the data has changed, the behavior changed. How do we couple with this in real life? Sure. So I'll answer the second question first. When you build a model, you can actually attach to the model the ability to self tune. Now the ability to self tune can take into account one of the models we did. We developed the model and then COVID started. And the ability for the model to actually self tune, help the model to stay accurate even after the changes in behavior because you have the COVID, for example, moved most of the customers to do transactions on the web and not to come physically. So it really changes the way customers are doing. And the model was able to actually take that into account. So you can have machine learning models do self tune. It doesn't mean that for years you cannot touch them on a periodic basic every six or nine months, depends where. You have to intervene and look at the model to see whether you need to add more features or take out some of the features. But you can have the model stay accurate automatically for the time it's working. And second, the main task in building a machine learning model, and this is why machine learning model cannot be a general model that will work everywhere or something you can buy and just implement, is actually preparing the data. I mean that's one of the biggest, most time consuming task. And even with so it all depends on the availability of data, of course. But usually, a model runs between three to four months. And if the second model you are doing, it depends on the same data that if you worked before, it would be much shorter. So it's not a few years in project and you can really see results fast, but it also depends on historical data that you have and others. And I would like to emphasize again something you said, Gadi, this doesn't replace the existing systems. Machine learning models and all this technology comes to augment what we have already today. A lot of the things we are doing today work best with rule based. Machine learning would not do a better job, sometimes would do a worse job. So a lot of what we're doing today is still valid, and we still need to do it. And even in five gs and IoT and what we will move to, a big chunk or a chunk of the controls will be rule based because they work great. But a lot of new things that we're doing, we to move technology. That's life. We cannot stay the same thing that we have. Now one of the things I talked about is not just the move of technology and the fact that we need to move to machine learning. It's the fact that a lot of the things that in machine learning were done before by batch at the end of the day, of the week, of the month, need to really move to real time. And fraud has been there, but machine learning revenue assurance has to move to real time. How can what can you tell us about how real time is playing here? Okay. Very interesting question. So as you said, fraud management have been real time for many years from its beginning. Revenue assurance and other areas of business assurance have not been real time, but this has to change. Why? If you recall in revenue assurance, we always speak about reactive revenue assurance, active revenue assurance and proactive revenue assurance. Reactive revenue assurance was about detecting things that already happened that you cannot correct. Active revenue assurance was about detecting things and correcting them before they impact the customer. So for example, if you have a build cycle and you detected an error, if you solve that error before it went to the customer, it was active revenue assurance, it was great. So you had the error, but the customer was not impacted by it. Nowadays, as you mentioned, we are moving to real time billing. We are moving to real time provisioning, self-service. And yes, in the past, we have some real time when we speak about prepaid, but now we are moving for everybody. And now if you just use traditional revenue assurance and you try to be active to prevent the problem before it impacts the customer, it will not work. Because the bill cycle is not at the end of the month. The bill will be or the charge will be in three seconds. So you have this very short amount of time to detect the problem, but not just to detect the problem, to correct it before the customer feels the problem. And this is very important. And this is really what changed the need for real time revenue assurance from something nice to have to a must. Because at the end of the day, the key things for the operators are the customer satisfaction. And if customers will see errors in their charging, will start to complain, It's a huge damage, so you have to prevent it. So you have to do real time revenue assurance and there are some challenges in doing real time revenue assurance because I remember many years ago people spoke with me about having real time revenue assurance. But at that time, first there was no real need for it. The second thing is to get the data about the things that are happening. It took time, so if you got the data into your revenue shown system, after three hours you cannot solve anything after three seconds. Nowadays, are we are moving to more and more standard APIs, it's possible to have revenue assurance as part of the consumer of these APIs to get the data in real time, to check things in real time and to correct things in real time. So to summarize, today we have the need, we have the technology and real time revenue assurance is not something nice to have, is to the core of the strategies of operators to move to real time billing and charging and to have happy customers. So in my view, it is a must and it's something that we can do nowadays. And can you elaborate a bit more? Doing something is real time is great, and and raising a flag in real time is great, but you need to do something. You know? Yes. One of the problems in the past was even if I get revenue streams in real time, what does it matter if nobody does with anything? So you can talk a bit about automation and how how we can actually get to the point that we're actually doing resolution in real time? Yes. So obviously, working real time means that you cannot just send a report to a person and ask that person to correct things. The person doesn't work in real time. So you need to incorporate to this robotic processes automation which will take all this data and will do the correct immediately in the system. So this is a must and it's part of the solution. And when you think about doing real time revenue assurance, you must incorporate into the solution this automation. Otherwise, it will be worthless. Okay. Thank you. Rezi, we spoke about emerging technologies. We mentioned things like cloud, marketplace, mobile or multi edge computing, Mac. How they impact Business Insurance? Okay. First, I want to talk about the fact that these things are impacting Business Assurance. There are also other things that are impacting, and Business Assurance scope is becoming bigger. If we talked about revenue assurance and fraud, which were the main pillars of Business Assurance in the past, Some of the things that are happening now are actually driving or expanding the the responsibility of business assurance managers. And if we talk for for example on cloud, one of the main things that cloud is doing is the fact that you need to migrate data from one place to another. And when you're migrating data, usually, you're also changing technology of systems and you're changing the type of databases that you're using. And and and that also changes some of the business rules that you're doing. Your billing system looks a little bit different and and things look a bit different. So this also now Business Assurance is also responsible to make sure that the migration went correct. So the data that I have on my customers on prem, if I move them to a system in the cloud, stay the right way. We have a lot of mergers and acquisitions now that have to move customers from one it's not just from one billing to another, it's from one company to another. And the billing system in the new company doesn't look exactly the same. So when I'm moving the customer, it's not a one to one migration, and I need to understand the logic in order to check that he because I don't want I want the customer to see a different name on the bill, but the rest stay the same. It doesn't need to have a new bill with the new amount in it and something that it doesn't understand. So moving to cloud is actually doing a lot of migration, and you need to assure the migration. So migration assurance is also part of business assurance today. And as we said, and also we are delivering the projects on cloud. So also the billing or the Business Assurance system, doesn't matter if it's a fraud system or revenue assurance or a bill QA system or other system are now operating on cloud. You need to have the system move to cloud and also changes the way you're doing your working because the information is already there. So you cannot keep it on prem if the information is moving to cloud. You talked about marketplace and the API economy. The the view to mainly five g, but also IoT and eSIM is that operators would not actually sell everything by themselves. They huge use huge amount of of third party partners that would actually sell services and solution on their platform. And I will bill it probably on your monthly bill or sometimes it will be closer to real time. But then I will have a lot of interactions with third party partners that would come and go all the time. It's not, you know, we used to have interconnect and and roaming. They were the same customers, the same contracts for years. It was easier to monitor. But today, it would change. Each one of them will have a different kind of business with us, a different way of splitting revenue, and and a different timeline of how what does he get. And I need to make sure that the customer got the service that he needs to pay for, that I got the payment from the customer, that I paid the third party what he needs. And there's a timing issue when I'm when am I actually collecting from the customer and then when am I paying the partner. So it's it's gonna be very complex, and we need to be ready for that, for some of it to also be in real time because some of the payments will be there. We we are also moving to an to an economy of of of transactions. If I had one bill until now, it looks very simple. I pay the same amount each month. I will be able to buy and add so many things that some of them, it will be on transaction based. Or revenue assurance need to check now transactions, not just at the end of the period. And there are also the distributed ledger. I will need to make sure my the expansion also goes to me making sure that everything is written correctly. Everything that I've done is written correctly in the ERP system because I will pay to the vendor through the ERP system. And I need to make sure that it's written correct there. So for example, one of the services we are providing is also something we call ERP assurance. Business assurance never really looked at ERP information. How are we handling with vendors which are not interconnect vendors? How are we paying expenses to employees? How are we looking at inventory, etcetera? We have to look at that also. It's a different system. It's not a revenue assurance, and it's not a fraud system. It's a different system looking at things differently. So we will have multiple business models of how we need to look at information and what are we looking at. And blockchain is coming in, and we need to understand exactly that works and what do we need to monitor and at what rate because this is also close to real time. And as you mentioned, there is the multi or mobile edge computing that will move our interaction with the customer from what we are accustomed to, to the CRM system or the billing system to the edge. And I need to understand how is this handling the customer now. And now am I making sure that all of the information is running correctly and everything is working correctly. So it really it would really change. Not everything is here today, okay? We need to be honest here. But it's coming, and it's coming very fast. And you don't want to be caught that your operator or your company is actually deploying it, and you don't really know how to monitor. You have to be ready. And as I said in the beginning, you have to know exactly what's to happen, and you have to understand how you need to do it, and you have to be ready. So you need to look at solutions today and not tomorrow. So Gadi, how do you summarize this? I mean we talked a lot. We talked about the changes. I know it's exciting times and everything is changing, and it's sometimes confusing to people our industry to understand exactly what they need to do. I mean there's so many change. We cannot do everything at once. But how do you summarize what are the next steps they need to take in order to be ready for everything that is changing? Okay. I think it's a really very important question. The first thing is that you need to get knowledge. Get knowledge first about what your company is planning. So if you are deploying five gs or if you are deploying a complex marketplace for over the top services or if you are moving to real time billing, you need to get knowledge about the plans of your company and you need at the same time to get knowledge about the new technologies and the new methodologies in business assurance. Things like real time revenue assurance, things like using machine learning and AI, things like using blockchain. You need to consider all these technologies. Once you have a view of all these, you need to set your goals. What I'm trying to achieve, reduce leakage, increase customer satisfaction, ensure that the transformation project will go correctly, that the customers will be happy after it, and define your strategy. And, Jesse, as you said, your strategy is not something that you can wait until you finish to deploy five g, until to you move to a new real time billing system. Because you are part of the success of moving to five g, part of the success of moving to the real time billing system. So you need to set the strategies, start to prepare for them, build the plan, see, okay, five gs will be deployed in three months, what I need to do in order to ensure that it will success. You need to find the partners because for many of these new strategies you will not have in house knowledge and experience. And you know it's very different if we speak for example about machine learning to have a person that is an expert in machine learning but it's newcomer to fraud and don't have really the experience of how to use machine learning to detect fraud in telcos, then to have people or companies that have already done this. Use machine learning in this area and this use cases. So you need to find the right partners to work with them and you need to, you know, as Nike said, just do it. Don't wait for everybody to do it. You need to do it in order to ensure your success. And it will requires effort from your part, you know, build new things, create new strategy, but you really have to start doing it. Thank you, Gary. You're right. It's it's everybody said, you know, it's exciting time. It's it's challenging. There are many challenges ahead, but there's also many opportunities, and we cannot avoid it. This is where our market is going. This is where technology is taking us. And as I said, it's not just technology. Looking at five gs, IoT and others as a technology change is not the way to look at it. It's a business change, and it's a huge I think it's one of the biggest. It's much bigger than what we experienced when moving to three gs or four or any other technologies, and it's also a combination, as we mentioned. It's those things are happening together. We're moving to cloud, we're moving to IoT, and we're moving to eSIM. And all of these things together will create a huge change. Thank you, Gadi, and thank you all. We'll move to some of your questions now. And I would also like to remind you that if you are looking to see more of what we're doing in machine learning, you can reach our website and you can see some of our success stories, especially with Bell that presented with us both in TM Forum and in the MWC, and you can see how they were able to solve some of their issues using machine learning on top of their fraud system that they have today. Well, Gadi, it seems that we have a few questions from the people online. First one is how long before we need to actually do the changes that we talked about. When is five g coming to our neighborhood? Well, five g is here, not in all of the territories, not in all geographies, but it's on its way. It's here. And we believe that you need to start acting now if you want to be ready for everything that is changing, everything that we've talked about. This is these are big changes that change not just what your company is doing, but also the way you should be actually doing all the controls on it. So we believe the time is now to actually look for systems, looks for consultancy, look for different processes and be ready because it's coming and it's coming now. Second question is, how long does it take to develop a machine learning use case? Okay. So from our experience, it typically takes between three to six months. Large part of this process is to bring the data, to understand the data, and to clean the data. So when you are developing a second and a third use case that use large part of the same data, you can reduce the timeline. It will be much faster. Great. Now another question is, can't existing platform handle the new risks? I mean, operators have existing platform either homegrown or from a provider. Why can't these platform actually continue to do the business assurance? So in my view, let's distinguish between existing platforms because the new platforms already exist and previous generation platforms. In my view, previous generation platforms can do part of their new risk, can handle part of the new risk. But they will will require much more work and it would be really difficult because we are speaking about applying machine learning to the new unknown unknowns. If you have a previous generation platform, your user, your experts will have to detect these unknowns and just then put the rules to the system. So it will take a lot of time. So yes, you can handle from the point of view of rules, part of the new risks, but it will be much more difficult. Okay. Another question is, should or does business assurance become part of the operational process? This business assurance for many, for a long time now has been a reactive process. I mean, things would happen, we would monitor it, we will raise problem, fix it all after the things have happened and never in real time. Now when we're moving to everything that we talk about now, real time is a key. A lot of the things will happen in real time. A lot of the charging will happen in real time. Customers will be able to see what happened in real time. So business assurance must become in real time And in some instances, like we talked about point of sale fraud prevention, you want to be part of the operational process. If this is a fraudster online now trying to buy and you know he's a fraudster, you don't want him to complete the transaction and then raise a flag that it's fraud. You want to stop the transaction from happening. So yes, business assurance in some cases, like in point of sale practices or sometimes in the billing process should be part of the operational process. And that means that, yes, we're becoming less of a reactive monitoring process within the operator and more of a real time assurance that is part of the operation. Not all of it. There's a lot that we'll be doing that'd be near real time or after. But yes, we are changing the role of business assurance within the telecom operator. That that that's the fact with the new technology that we're having today. And last question, Gadi, is really we talked about it but it comes again. So are we moving to full machine learning and no business rules in place? The short answer, no. So if you have rules, if you have knowledge knowledge about some risks, put this knowledge into rules. You don't need machine learning for that. On the other hand, and as we see using rules, using statistical analysis, there is a limit to what you can achieve. We see that there are still large amounts of revenue leakages, large amounts of fraud. Machine learning can help you to close the gap, to build on top of your rules, on top of your statistical analysis, and reduce the amount of leakages and of fraud. Okay. To summarize everything we've talked about, I mean, let's not neglect what we're doing today. There's a lot of what we're doing today which is still very, very relevant even in the future. But we need to be ready for the new things that are happening. We need to be ready for new technologies. We need to be ready for our new business assurance role within the company. So thank you very much for joining us. Again, for more information, you can connect to our website. See in a day or two after we finish doing that, you'll see the full webinar again and more information about business assurance in five gs. Thank you very much. Thank you. Thank you, Dadi. Amazing. Make it it Okay. I'll kiss head off for nothing but amazing. The future is waiting for my eyes. Wake up in the morning, riding up the highway, walk through the door light, we put the a in. Okay. So last year, no game if you're less than, because so so should not go where we come from.</p></div></div></wistia-player>