PegaWorld iNspire 2024: Pega AI-powered Transformation: NTT East's Next-level Automation

All right. We're going to go ahead and get started. So I have the honor and privilege of introducing some folks from Japan. And as we've worked together to prepare for PegaWorld, I've really seen the passion that they have for the digital transformation and the work that they're doing with Pega. And so they've put a lot of time and energy in taking that passion in Japanese and putting it into English. So they're working very hard to really tell this story today. Feel free to just capture your questions. And at the end we'll have a Q&A. So I would like to introduce first Uki Masa Mimura and Udaya Komiya.

And they're going to be sharing with you today their journey of transformation using automation. And what's really good about this story is like we speak at Pega. Change happens small and over time, and so they've been able to build on some very complex order management that they've put in place and built out from RPA into digital messaging and into other technologies like automation and AI chat. And they're excited to tell you and show you how they're going to even expand that now into Gen AI. And we're excited about that. Based on everything that you're hearing about with Gen AI, and they're going to be at the forefront of leading that with Pega. So without further ado, please invite them to the stage. And thank you for being here. With you Scott.

Thank you. Hello, everyone. Nice to meet you. We are from Japan and we work at NTT is we are here to present a case study of next generation business automation at entities using Pega AI. This is happy traditional costume worn at festival in Japan. And this sandals is called Zeta together happy and Zeta make up perfect outfit. Wait, wait. A happy coat and Zeta are not a perfect costume. You need this tenugui.

Oh. Thank you. This is perfect. How do you like it? Does it fit? Las Vegas? Let me introduce my myself a little. My name is Yukimasa Mimura. I am project management of system development.

I have been working at NTT East for about 30 years. I am excited to be here today to present our initiatives. Hello, this is Udaykumar. Like Mr. Mimura, also from Japan. I am 24 years old and the youngest speaker at PegaWorld. This year I joined NTT East last year and have been participating in project while learning about project management. Here are four things I would like to share with you today. First of all, I would like to explain about entity, group and entity.

Entity Group was established in 1870. 150 years ago. Was still a time when samurai were still walking around with Japanese swords. Back then, we started out in telephone business. Since then, entity Group has provided high quality and stable telecommunications infrastructure as an information and communication provider, including the provision of broadband access service using optical fiber. In recent years, we have worked to solve local issues and create values as a formula ICT company. This slide aims to convey the size of the entity is based on the number of employees and operating revenue. Business activities include regional telecommunication business in the East Japan region. Business activities incidental to this business.

Business activities to achieve the company's objectives and business activities to utilize the company's resources. The registered capital was 2.2 billion USD and employees were increased to 4900. Also, our consolidated operating revenue was 11 billion USD. This is the sixth largest operating revenue in Japan, after Toyota and Honda. Our headquarter is Shinjuku, Tokyo. We are within walking distance of Shinjuku. Kabukicho. This busiest shopping district in Japan. And the famous Kaminarimon in Asakusa is nearby.

If you are ever in Japan, please let us know and we would be happy to show you around Tokyo. Next, we will explain the characteristics of our business and what we have achieved with Pega. NTT East receives approximately 8.5 million orders for internet lines per year. Workload. For internet and telephone service a customer places an order via the web or over telephone. The back office feeds the information into the back end system and the customer receives an invoice. One operational characteristics is that there is only one back office for multiple sales channel. At the time, the back office was operating in efficiently for orders from the sales channel. Specifically, they include the following for first checking for incompleteness.

Second, confirmation to the sales channel by phone or email. Third manual system input. Fourth paper based operations. Therefore, we considered automating system input work in the back office. The actual realized form is shown below. Pega Cloud and VA were linked to automate the input process. This eliminated inefficient tasks such as checking for implements and paper management, and provides the efficiency of back office operations. Effect of the introduction. Linking Pega and RPA not only improved operational efficiency, but also enhanced security.

Specifically, issues related to implementing incomplete applications were eliminated. The communication with sales channels, which used to be conducted by phone or email, has been converted to chat and paper reduction has been realized as of April 2020, for nearly 1200 tasks have been automated, which is a great achievement. The method is introduced here as an ingenious point related to result of this project. In order to automate operations, it would have been very inefficient to design and develop workflow one by one. Whenever there was a workflow to be implemented, it was simply put into one of those patterns and registered in Pega Cloud. If there was work to be performed, it was simply placed in those patterns and registered as Pega Cloud. As a result, 120 tasks were automated within three months of implementation, and as many as 1200 tasks are currently automated. As we have mentioned, we have successfully automating many tasks by ranking Pega Cloud and RPA. In particular, we have achieved great results in impaired work and even now we we are expanding the number of tasks to be automated every month using the Katagami method.

We will introduce the DCS more deeply. I would like to explain the shocking meeting. In a comic like style, just like Japan. And left side is the left side is side, Mr. Webb. Automated. Automated. The input process. And we're having an easy time.

See, I am completely relaxed. Then Mr. Komiya said to me. Chief, may I have a word? I await his word to see. What a happy report. I will hear. Of course, at this time we still do not know about the new issue, he said. Actually back office.

Is it not an input process? I keep thinking about his report. It's not just an input process. I didn't notice. I cannot believe there is more to the back of it than typing and quite a few of them. We thought that if we could improve the efficiency of the input process, we could automate the back office. Once again, we have organized the elements for improving back office efficiency. The first is input process. This is automated by Katakami method we explained earlier.

The second is to respond to inquiries. We have various inquiries or on the bucket for example checking incomplete orders and responding to inquiries. After receiving an order without automating these tasks, we will not be able to improve the efficiency of backend operations. Another important factor is intuition and experience. This refers to check and responses to inquiry that only operators who have been performing backend operations for a long time would notice. This is an issue that will be addressed in the future. The back office receives many inquiries from the sales channel, such as checking the inventory of goods as well as from related departments where where orders are issued. We found that we respond to 80 million inquiries per year. Most of the inquiries are still conducted by phone or email.

Specific inquiries were categorized are confirmed as before and after orders were issued. Various types inquiries, such as checking inventory and handling confirmations after application. We therefore decided to solve the problem by selecting inquiries related to inventory confirmation and checking entry details from among the many inquiries we received. This is a specific inventory check workflow. The Excel forms generated by the sales channel are routed through the back office to the goods manager, who then received the responses. The point is that the back office cannot check inventory in response to a request from the sales channel, so it turns over to the goods manager. Inventory checks are performed by the goods manager, but since they do not respond directly, the gut check response is received by the back office. As you can see, the back office was only responsible for the back office via forms and was conducting very inefficient operations. Therefore, we decided to use Pega, the instant messaging service, to handle all the tasks that have been performed by the back office and the goods manager.

Specifically, three tasks task performed by the backend and good manager. One request for inventory to conformation inventory. Three conformation reply. These three operations are realized by the DMs. With the DMs, inventory checks are completed at the sales channel, eliminating the back office and goods manager tasks that were previously performed. In addition, Excel forms that were previously used as request forms have been eliminated and operational efficiency has been improved in the sales channel as well. We will now give you a demonstration of the inventory inquiry. There are two key points to the demonstration. Active, appropriate scenarios.

Response to variations in operator skills. The second is response speed from input to response. Now let's do a demo of the invented inquiry. Let's start with this icon. The chatbot will start up and a message will appear saying welcome to the chatbot. In this case we type I want to do this talk to inquiry about inventory. Here's a point. For example, if you enter something like, I'd like to introduce our inventory of our inventory. Inventory inquiry scenarios will be launched.

This is a function unique to DMs. If you enter an item code, the corresponding inventory is instantly displayed. You can see that the completed workflow of filling out forms in Excel, sending them to the back end and receiving a response a few hours later has been made more efficient. We will further explain the Azure As before, we are launching and proceeding with the inventory check scenario. Here is a point. In the previous scenario, we had to enter an item code to determine the item, but until now, the operator did not understand all the item codes and had to search for the item code in the manual. Now, if we enter an keyword such as Wi-Fi, the system will sort and display the appropriate equipment, or the operator had to do is select the appropriate item to check the inventory. We also worked on requests for goods. This is a workflow before the introduction of DMs.

As you can see here, requests are made using Excel forms and various tasks are performed in the back office. Here too, DMs was installed in the back office to automate each task. Here is a flow after the introduction of the system. Although some of delivery checks need to be done by human eyes and therefore have not been automated, you can see the significant reduction in the amount of work that has been performed up to this point. This will be a demonstration of Over request for goods. There are two key points here. First, as before, we are minimizing the number of input items for the 4000 operators who use the system. Second, the information obtained by the AI chatbot is linked with RPA to automate system input. We will began demonstration demanding goods.

Click here to start. Now type in. I want to request. I want to request. Yes. This will launch the goods request scenario. From here on, the chatbot will listen to everything you say name ID, address, etc.. Here is a point to minimize operator input. We've made it selective.

From this point on, just answer the chatbot question as it asks you. Once all the information has been entered, the system will automatically input it into the system. There are three things we have achieved through this initiative. First of all, the chatbot has improved the efficiency of inquiry operations as shown in the demo you just saw. The response time is so fast that inventory can be checked even while the operator is still dealing with the customer. Furthermore, by linking AI, chatbot and RPA, the entire workflow from inquiries to input into the core system has been successfully automated. We are told that this is a world first case of solution linking Pega, DMs and RPA. Third, we have achieved zero back pain in this work. We have been improving the efficiency of input operations, but for the first time we have been able to improve the efficiency of inquiry operations and reduce them of zero in the future.

We would like to expand the target business by developing this platform to other business operations. Here are some reflections on this project. As you can see, only inventory checks and request operations are automated. It is difficult to use AI chatbot to handle all inquiry tasks, because back end operations do not include only simple inquiries like inventory operations. There are some tasks that cannot be answered without checking the manual, such as service fees. In the future, it is likely to become a new challenge to respond to these tasks. This is exactly the kind of work that intuition and experience that I mentioned earlier. I will explain our plans for the future. This slide describes what we have implemented so far and what we would like to try in the future.

On the vertical axis, back office operations are listed by difficulty level. The horizontal axis shows solutions that utilize Pega. Since we have achieved automation for input, confirmation and simple inquiry tasks, we would like to tackle complex inquiry tasks in the future and to achieve process automation by using Pega Knowledge Buddy. Here is an image of Knowledge Buddy in use. There are many complex inquiries in the back office that can only be answered by checking manuals and other documents. When we receive complex inquiries, we have to search for manuals on the web or on paper on the back end. We intend to utilize Knowledge Buddy functionality to Change Manager and respond to these inquiries on behalf of human operators. That's all. Finally, I just want to say thank you for this project.

At the end of the movie. Excuse me for speaking Japanese. Sorry. Pega project. NTT data. Hajime. To. Support the project. Buenos.

Aires. Is. That all? Thank you very much, sir. Open up to a session. Uh, do you have any questions or comments for the presentation? How did the architecture of Pega speed up the amount of processes that were available for such a short time? Um. The.

Cycle continues. To. Maybank. Maybank. Pega. So when it comes to system architect. Practices consistency. Is ecosystem. So every morning uh Pega can ingest, um, inventory system data.

So Pega can store the data so it improves performance. Any other questions? How do you find the language support with the bot in Japanese working? For. You. Are you asking about DMs? Digital messaging service. Okay. What do you use to install the system?

Uh, so Pega DMs is incredibly good, uh, when it comes to performance. And. Uh, 96% of like. Oh, okay. Um, 96% of, uh, DMs can answer the questions or inquiries at 96% with right answers to back to you. Yeah. Sorry. So the rest of 3%, uh, we, uh, we use machine learning to improve quality. You have to and you have to.

Change any. Back office. Processes so that you can automate them. You know. The process we use. So we don't change any business processes because we adopt DMs service. So there's no change. Uh, I think. Yes.

On behalf of our guest, Christian, uh, from Toyota Finance. The question is about what RPA has been used. The first question is this Pega RPA or the RPA? Yeah, we we. We use win Acta. OK. Uh, that NTT solution. Okay. And the second question is, is there any impact of the upgrade of Pega Platform connecting with RPA?

RPA impact. System to the. Network from. The workflow. We have two different network networks, but when we adopt Pega it, it connects each other so that the outcome. Any other questions or comments? All right. Well thank you all. That was great.

I appreciate all of your passion. Thanks for attending and see them and have any other questions throughout the event. Please feel free to ask. So thank you for coming. Thank you.