Hello to everyone watching. Thank you very much for joining us for SHIFT's Q2 broadcast for the fiscal year ending August 2026. I am Masaru Tange, President and CEO of SHIFT. In today's briefing, I will report on our Q2 results, focusing on SHIFT's AI initiatives and incorporating answers to questions we frequently receive from our investors. We will have a Q&A session after my presentation. If you have any questions, please enter your organization's name and your own name, and then type your question into the chat. Also, we would appreciate your cooperation in submitting your questions by the time our presentation concludes. Now, let's begin the presentation, h ere is the agenda. First, we have the performance summary. Regarding these results, we recorded record-high net sales and gross profit. We have been investing ahead in AI, and we expect to see a return on that in the second half, which I will discuss later. As for the key points regarding net sales, even on a non-consolidated basis, we saw 122% growth so sales are growing quite steadily. Also, the gross margin was 34.4%, which is a recovery of 1.9 points compared to Q1. In addition, SG&A expenses increased by 2.3 points- 23.4%, but we invested a total of about JPY 2.5 billion, and we have a firm prospect of recovery. Moving on to Q3, sales are trending steadily toward exceeding 120% growth, and we anticipate they will be around JPY 40.5 billion. Also, regarding the gross margin, we are currently working hard on it, believing we can reach about 35%. As for SG&A expenses, since the investment in AI has settled down, they will decrease to around 22.5%, and we believe we can achieve an operating profit exceeding JPY 5 billion. Next is the details of our AI investment. This fiscal year, over the first half, we have been investing in AI for the past 6 months. As you all know, AI has become a major global topic over the last 6 months and has gained significant power. Naturally, we intend to utilize AI to firmly grow our business performance. Combining both cost of sales and SG&A expenses, we have invested about JPY 2.5 billion. If we subtract this investment portion after deducting AI costs, sales have naturally grown by 116.8% and regarding operating profit, it's at 117.3%, so I believe we're seeing solid growth in areas other than AI. Next, I'll move on to our AI strategy. I've been receiving many inquiries from investors lately. We are frequently asked whether AI will eventually replace traditional software testing and development processes that involve human expertise. We believe these areas will be redefined by AI, and our role must evolve accordingly. Along with that, a second common question is whether SHIFT's work will vanish as companies move toward in-house production as AI advances. Regarding this, SHIFT itself uses AI natively, so we've been effectively streamlining our operations. In that sense, companies that can master AI will naturally move toward in-house production. However, many companies still struggle to master AI or use it to boost productivity, which is quite common. I believe they'll continue to outsource, and since we're ahead in building various AI services, which I'll explain later, I think they'll utilize those services. Now, I know there's been concern over the past 6 months about whether SHIFT will be replaced by AI, so I'd like to explain our response to that. Over the past 6 months, or looking back about 3 years, we've been working on our AI initiatives, I believe SHIFT has finally struck gold with its AI strategy. Looking at the global context, AI investment is currently focused on things like semiconductors, data centers, and power companies. I believe a revolution is starting from that kind of infrastructure. While those areas will continue to grow, now that the LLM competition has stabilized, I believe the focus of AI competition is shifting to the application layer. Secondly, while companies worldwide are trying to build services or boost profits using AI at this application layer, many are stuck in the POC stage and failing to see results. We've developed and secured orders for a service that generates solid sales and profit with high gross margin, so I believe we've struck gold ahead of others in Japan. This is the outline of our AI initiatives. We've been working on various initiatives, not just in testing, but also in development and consulting. Going back, our efforts in AI began in 2015, which was over 10 years ago now. As you all know, AlphaGo made huge headlines when it defeated a world chess champion. When I heard that news, I thought we could build something with AI, so independently, even though we hadn't mastered AI yet, we built our own system. We created a testing service and device that automatically reads systems and used robots to perform operations via key striking machines. Secondly, while we built those services back then, as time passed around 2022, we realized AI would be vital. We brought a Kyoto University AI startup into the SHIFT group to gain the technical power to build our own LLMs, acquiring that tech 4 years ago. Last year, exactly 1 year ago, we made our AI native company declaration. Since LLM performance has improved dramatically, I felt we absolutely had to utilize it. From that point, SHIFT began building a variety of services that led to our AI-driven modernization service. Since July 2023, over about two and a half years, we have created 2,395 AI agents. Furthermore, we've focused heavily on this over the last 6 months, investing JPY 2.5 billion in AI in just that short time. This has led to the development of a truly transformative service. It leverages AI to automatically analyze source code and generate design documentation, I'll explain the detail later. But in just 2 months since its release, we've processed 100 million lines of code, creating JPY 30 billion in development opportunities. Actually, this is the important part, but through that service, we visualize source code and then use AI to auto-generate systems. In just 2 months, we achieved JPY 2.55 billion in sales with a 60% gross margin. I'll elaborate later, but we believe the target market for this is JPY 7.5 trillion. I think this is just the first chapter in capturing that JPY 7.5 trillion market. I believe AI will disrupt many industries, so I've researched and compared things from a broad perspective to see how much our own business might be disrupted. I'm sure you all experienced the era when feature phones shifted to smartphones. Furthermore, from horse-drawn carriages to automobiles, CDs to streaming, and film to digital cameras, various industries have been transformed by new technologies. I believe that history has shown us how these industries have been replaced over time. Specifically, we believe that the disruption from feature phones to smartphones occurred in about 7 years. What's characteristic about this is that from the era when hundreds of millions of feature phones were sold, it went beyond that. Smartphones, like iPhone, began to sell even more. It wasn't just about selling hardware. Software also started to be sold. By software, I mean applications, t hey started selling those as well. The feature phone market was only JPY 13 trillion, but now smartphones sell JPY 70 trillion as hardware, and the applications on them have reached JPY 83 trillion in sales, which I find to be a very interesting development. What's noteworthy here is that even when smartphones appeared, feature phone sales initially continued to grow. Eventually, they were phased out and sales declined, but a certain number still remain, which I think is a key characteristic. Based on this historical context, in the IT industry, we use the term AI native. We believe that new businesses and services fully utilizing AI will replace existing ones over a period of about 10 years. Based on that background, this graph shows what kind of strategy SHIFT should pursue. What I'm explaining here is sales on the vertical axis and time on the horizontal axis. We have provided various services like system construction, testing, and consulting based on a man-month model. I think this corresponds to the feature phone era. I believe the JPY 16 trillion market will initially grow. AI won't spread instantly, and the global labor shortage remains a reality, so it will increase temporarily. However, it won't keep growing forever a s AI services become more prevalent, it will likely decrease, but it won't hit zero. Now, regarding the gross margin, we currently have a 35% margin. I believe that man-month services will continue to yield that kind of gross profit. After that, I believe the next phase will be the transition from with AI- AI native. By with AI, we mean leveraging AI to enhance and build upon existing solutions. I'll explain later how we've built services for both with AI and AI native. What's particularly disruptive is the AI-driven modernization service I mentioned earlier, which automatically analyzes source code to create design documents and then generates source code to build systems. I believe this will be the core service in the AI native era. At that point, we should be able to achieve a 60% gross margin. Since we're already at 60%, as AI performance improves or our technical expertise in using it grows, 70% or even 80% is not just a dream. Plus, from the client's perspective, buying a system isn't everything. Performing and processing business operations is first and foremost the primary problem-solving goal for clients, I believe. In that sense, I think the market for BPO, which handles business operations, and BPaaS, which includes it, will grow from now on. We call it AI BPaaS, and we intend to firmly build an outsourcing framework using AI agents. I believe this will become the equivalent of the app market in the smartphone era. We've waited 10 years for this. Instead of just providing man-month services like consulting, development, and testing, we want to build it thoroughly from this stage. In the coming AI native era, we aim to create those systems, services, and business models faster than any other company or industry. 2026, I think this is the inaugural year, but as I explained earlier, the service is already ready. We're generating revenue and achieving solid gross margins. It's in high demand, and we're constantly receiving orders from clients. I wonder how far this will expand. Man-month development, consulting, and testing won't disappear immediately. Keeping that in mind, we want to properly manage resource adjustments and reskilling. To summarize, no matter what era comes, whether AI arrives early or its full impact is delayed, we have successfully transitioned to a flexible and resilient business model, no matter what changes may come our way. Now, if you take a look at this diagram, it will help to explain in much more detail exactly what that new state will look like and how it will function moving forward. We've divided it into operations and systems, but within operations, we believe there are non-core and core tasks. For example, systems for managing operations like Salesforce, Oracle, and Workday are already widespread globally. Whether these will be suddenly replaced, there's talk about UI and UX changing, but since they have solid databases, if those databases are AI-compatible, they'll use them for AI, so I don't think they'll disappear easily. However, on top of that, there are business rules and workflows. For instance, tasks like sales administration, accounting, and labor management are being performed. For these parts, which vary by company, I think they're closer to non-core than core operations. I believe those areas will proliferate by using general SaaS products with AI agents layered on top, SHIFT will target that as well. On the other hand, for core operations, services, businesses, or systems essential for a company's revenue or differentiation, general SaaS or package products aren't enough. I believe a world of building original systems from scratch will open up. In our company, we develop our own systems for core business operations, such as our proprietary project management system. A marketplace for managing business partners, and our well-known personnel evaluation system. We build these core systems in-house. As I mentioned earlier, we offer an AI-driven modernization service. We aim to support our clients by building these systems quickly, affordably, and effectively. To make it easier to understand, I've divided SHIFT's AI strategy into four segments for this presentation. The 1st is Segment Zero, which covers our traditional consulting, development, and testing services, t his accounts for 88% of our revenue. Next is Segment One with AI, w e believe there's a clear demand for using AI to improve the efficiency of consulting, development, and testing. Specifically, tasks that used to require 10 people can now be done by 7 or even 5 people using AI. This segment already accounts for about 11% of our revenue, and we've built it to achieve a gross margin exceeding 40%. Expanding this with AI business and its services is a key focus for us. Furthermore, in the medium to long term, expanding the scope of AI native has become extremely important. As I've emphasized from the start, we've already established our AI-driven Modernization Service. We're operating on the premise of steadily expanding this, and we aim to capture direct client-facing business through what we call AI BPaaS. First, let me go into more detail about our AI-driven modernization service in the era of AI native. In the past, and I've been in this industry for over 20 years, this has always been the situation. When major system integrators or consulting firms are asked by clients to modernize legacy systems because they're unmaintainable or they want to replace them or build new services, this is how they typically handled it. Basically, design documents are nonexistent, a ll that's left is the source code, and that source code is filled with comments. For some reason, the business logic within an if statement is explained in plain Japanese. Without those comments, engineers reading the code later wouldn't have a clue what the logic was supposed to do. They'd have to manually check and decipher cryptic source code, sometimes millions of lines of it. Of course, that's not enough, so they'd interview engineers on-site, try to extract tacit knowledge buried in industry expertise, or refer to outdated design documents for any clues. That's how it's been done, and it's still happening today. Roughly speaking, if you have a system with 1 million lines of code, it would cost about JPY 1 billion to build. Now, visualizing the documentation for that is something consultants often do, and we're currently taking orders based on estimates of about JPY 50 million over 3 months to get it done. It doesn't always end there. It's quite common for projects to stretch from 3-6 months, ending up costing JPY 100 million. As I've mentioned repeatedly, for that analysis itself, we use an approach called DQS, and we automatically analyze the source code using AI. This allows what used to cost JPY 100 million to be done for just JPY 10 million. Furthermore, what used to take 6 months can now be generated in just 2 weeks. Plus, since it's not made by humans, 46 types of system design documents are automatically generated by AI. So there are no mistakes and all business logic, including which parts are connected to the database and how the business logic is processed, is generated in a well-documented form. This diagram shows the types of business that will stem from this. We call this service DQS, and DQS Reverse is the service I just explained. This enables up to a 90% reduction in costs, allowing us to deliver solutions at significantly lower cost to our clients. After that visualization, what we do next is a process called strategy or ST, where top-tier engineers develop the concept for the ideal system, which is the architectural planning shown here. Based on this concept, we move to what we call forward, which is the actual system development phase. As a result, we can now deliver forward engineering to our clients at approximately 30% lower cost. Also, and this is quite interesting, by performing reverse engineering, we can understand all the causal relationships. For example, we can identify potential regressions or where system failures might occur, and when adding new features, we can determine exactly where to integrate them so the system continues to operate normally. We provide operational support while keeping these factors in mind, and since everything is visualized, we can offer it to clients with confidence. In short, we've created a service that allows clients to suddenly switch to us from the maintenance and operations currently handled by existing SIers. Generally, maintenance costs are said to be around 20% of system construction costs, whereas we can provide the same for approximately 10%, representing about a 50% cost reduction for our clients. For all these services, we can provide them with a minimum gross margin of 60%. We've built a service that is cheaper for the client while increasing our own gross profit. Moving on to the details. Over the past 2 months, we've been actively securing orders through reverse engineering within our AI-driven modernization service. As I mentioned earlier, we've taken on 100 million lines of source code, representing a JPY 300 billion opportunity. Out of that, we've already secured JPY 25.5 billion in orders in just two months. To break it down, we've done 53 free reverse engineering projects, 14 of which converted to paid ones. There were 6 strategy and system planning projects, and we've actually built 16 systems totaling JPY 23.6 billion. There's clearly a strong demand from clients for rapid development, so building systems with AI seems to be very popular. Also, to reiterate, the maintenance and operation phase for ongoing support is something we'll be taking orders for moving forward. As I said, we think this is a prime opportunity, so please look forward to that. Next, we have some client feedback. Since we released this service 2 months ago, we've had a great response from about 6- 700 companies who have inquired about it. I've listed some of the feedback our sales team received here. What was particularly interesting was that clients really valued how we could visualize such detailed business logic and causal relationships. They were also amazed at how quickly we could produce results. As I mentioned, if you ask a major consulting firm or SIer, it might take over a year, but we can visualize a system in just 2 weeks. I think this was a truly shocking service for our clients. Furthermore, after the actual visualization, regarding client feedback, we've heard that project costs could be cut by more than half. Also, when we visualized their systems, we discovered unrelated HR systems were intertwined, which was quite a surprise. In one case, a client asked for a quote for 200 screens, but our investigation revealed there were actually about 1,000. It seems many clients don't fully grasp their own system scale. Next, this page shows where the opportunities lie based on our JPY 2.5 billion in orders. Surprisingly, we've secured many public sector projects. I wondered why, and it turns out that in the public sector, factors like technical skill, cost, and industry knowledge are clearly quantified in competitive bids. In other words, instead of just relying on traditional system integrators, the best company is chosen from a completely level playing field. We've won every single one of these competitive bids so far. Even against major consulting firms and system integrators, we're winning projects worth over JPY 1 billion. What's particularly interesting is that we're winning because our technical level is the same, but our costs are significantly lower. Also, regarding the lead time for these orders, if there's a competition for a public project, we can basically win it, allowing us to secure work very quickly. We plan to build our track record, brand, and sales here first. The market is huge, around JPY 2 trillion, so we want to build a solid foundation. Also, with this type of service, existing vendors often have a lock-in, and projects tend to go to companies with long-standing relationships. To address this, we're targeting small and medium-sized enterprises to introduce our service. In reality, major system integrators often can't handle these smaller companies, or their prices are just too high. Also, there's a lack of technically skilled integrators in regional areas. These companies want to outsource, but they struggle to find anyone who will accept their budget. However, for these small and medium enterprises, just as it was for us, when it comes to visualizing operations or building systems, our fees were naturally high, so we couldn't take on these projects before. Thanks to AI, those fees have dropped dramatically. We can now scale using AI without needing a large team, which I believe has opened up this small and medium enterprise market. The lead time for these orders is about 2-4 months, and while the project size is small, from the client's perspective, we're stepping in where no other company will. Everything is visualized, providing peace of mind for future maintenance and operations. For SHIFT, this yields a gross profit of over 60%, making it a highly attractive market. We've estimated the market size to be around JPY 2.5 trillion, so we're committed to building a solid track record here. Since it's about JPY 5 trillion, we're determined to secure projects in this area. Lastly, as I've mentioned repeatedly, regarding maintenance projects, the lead time for orders tends to be long. However, as I said before, maintenance usually accounts for 20% of development costs. Being able to do it for 10% is very appealing. Even if we take on these projects in the future, we'll be operating systems with a clear understanding of the scope of impact. We estimate this market to be around JPY 3 trillion in total. When you combine everything, we're looking at a market of about JPY 7.5 trillion. Within just 2 months, we've already secured JPY 2.5 billion in orders from a portion of that. Next, I'll move on to explaining AI BPaaS in the era of AI native. This is structurally similar to the AI-driven modernization service, but since we need to visualize business operations, we'll start with visualization. As some of you may know, the traditional way to visualize workflows or create manuals involved sifting through large amount of data. If no workflow existed, it meant interviewing individuals, whether temporary staff or employees, who held all that specialized knowledge in their heads. As noted here, this often involves paying high consulting fees, like JPY 4 million or JPY 2.5 million over 3 months just for visualization. It's common for projects to end up costing around JPY 20 million. For instance, a job might drag on for 4 months and end up costing JPY 30 million. That's a typical scenario. If we can get the inquiry data, we can automate the visualization to some extent. We've built a system that completes in just 2 weeks what used to take four months. Furthermore, a cost of JPY 30 million can be reduced to just JPY 3 million. We believe we've created an excellent service. Similar to the AI-driven modernization service, this reverse operation allows us to divide the service into 3 parts, the strategy for redesigning operations, the forward operation for actual implementation, and the maintenance for ongoing support. From the client's perspective, the discount amounts would be 90% for reverse operations and about 30% for strategy. Also, for forward operations, like building businesses and workflows, we can reduce costs by about 30%. For maintenance or business operations, we believe we can offer services at about 50% less. We also expect to achieve a gross margin of around 60% ourselves. However, regarding BPaaS, we aren't quite at the full-scale order stage yet. Since it's a very new service, we're conducting sales activities in parallel, though we have received some orders. As I mentioned, we aren't seeing the same level of results as our AI-driven modernization service yet, so we want to focus on awareness. To that end, I believe the first thing we must do is achieve solid results within SHIFT itself. SHIFT has approximately 15,000 employees, of whom around 1,500 are engaged in back-office functions. As a first step, we are visualizing and analyzing their tasks and progressively automating them using AI. As a second step, given that we have around 40 group companies, we plan to centralize these back-office operations into a shared services center and further enhance them through AI within this fiscal year. Once we achieve that, we'll have surplus personnel, and we want to use those members to take on more BPaaS work. We're actively selling to clients and are currently in the middle of an initiative to transform the back office from a cost center into a profit center. Now, here's an example of our AI initiatives at SHIFT. As I mentioned, we broke down the corporate tasks involving about 1,500 people over the past month. In just 1 month of analysis, we identified 892 distinct tasks. Also, and this is very interesting, we felt it was crucial to set actual unit costs for these tasks, so we calculated the processing unit cost for every single one. If you can imagine, for tasks like sales administration or order processing, SHIFT handles about 12,000 orders a month, which requires about 4 full-time equivalents. Now, SHIFT's revenue is currently JPY 150 billion, but if it grows to JPY 300 billion, that would require double the staff. If temporary staff are doing that work, the quality can fluctuate depending on their individual skills. Furthermore, if a temporary or permanent employee leaves, you have to find a replacement and train them. Since knowledge often becomes siloed within one person's head, their departure is a major risk because it can bring operations to a complete standstill. In that sense, for this order processing work itself, we've visualized the processing cost per case. By building it with AI, the current processing cost for order processing is JPY 552. We believe this can probably be done for about half that. If we can achieve this, even if sales double, for example, we'll have a system where we can process orders without increasing headcount. The unit processing cost is very important, and we've started by trying to establish that. In response to that, we've turned all back-office operations into AI agents. We've created 2,395 of these, and they're now being applied and are actually operational. In just 1 month, as of April 14th, two types of processes have become fully operational. These are order registration and quote creation and billing. For these two tasks, we've been able to reduce the workload by the equivalent of 6 people. We plan to expand these actual initiatives at SHIFT by this August, aiming to reduce the workload of full-time employees by 100 people and the number of business partner staff by 110. This AI BPaaS isn't just an internal SHIFT initiative, w e've actually already started providing it to clients. Here's what clients say the moment they see this visualization service. We often hear, "It's amazing to see this much from this data," and I'm very happy when clients tell us, "I can see a way out of being treated as a cost center." Also, when we show the AI agents we've built to clients, they say things like, "It's revolutionary to be able to use these AI agents without changing our current methods." Since we used them extensively ourselves, we receive many comments from people who want to emulate our results based on that track record. Now, let's talk about how we view the market for AI BPaaS, which stands for Artificial Intelligence Business Process as a Service. Understanding this market is important because it gives us insight into the opportunities and challenges that are shaping the industry today. First, as I've mentioned, we want to focus on building a solid track record within SHIFT. Next, regarding the public sector, since these are competitive bids, we can win them even without a big brand name if we have strong technical capabilities or cost competitiveness. We want to apply this AI BPaaS to those areas, w e're already winning public sector projects, even if they're not full AI BPaaS yet. We can see that if we can integrate our technology there, gross margin will improve significantly. Also, I believe this will be welcomed by small and medium-sized enterprises, as well as for operational tasks. In my view, there's a market for replacing temporary staff with AI agents. In cases where companies are hiring because of labor shortages, they can bring those tasks in-house by creating AI agents instead of hiring, I believe we can help in those areas. Next, I'll report on our business in the with AI era. Regarding with AI, we're currently working on various projects. Our top-line revenue has grown by 123%. For the Q3, we expect to exceed the Q2 and achieve a solid revenue growth rate. Also, our various KPIs are showing steady growth. What's important here are the unit prices per project, per person in charge, and per client, as well as the number of personnel assigned to them, o ur activities are becoming well-established. The number of persons in charge per company is increasing, and the unit price per project is also growing. As a result, the unit price per client has also seen significant growth. Regarding the repeat rate I mentioned before, which is the increase rate minus the decrease rate, we've successfully raised the increase rate while keeping the decrease rate at a low level. This shows the scale of our with AI revenue. In the Q2, we achieved about JPY 4 billion in with AI revenue. For the full year, we expect to reach about JPY 16 billion, which is about 10% of the total. Our overall business and with AI gross profit are shown here. We're focusing on development and testing, and in development, we've seen a 5.5 point improvement. In testing, we've also seen a 3.4 point increase in gross profit. While continuing our focus on development and testing, we also aim to improve gross profit in areas like consulting and BPaaS. Finally, to summarize, we've reported on various aspects of the first half, and we're committed to generating solid revenue in the second half. Our initial forecast was JPY 150 billion in revenue and about JPY 20 billion in adjusted operating profit. Thanks to M&A, we now have JPY 160 billion in revenue in our sights. We're also determined to reach our JPY 20 billion operating profit goal. By increasing gross profit and using AI to control SG&A and recruitment costs throughout the second half, we believe we can successfully achieve this full-year scenario. Lastly, as we work toward our JPY 160 billion revenue goal, we expect our workforce composition to change like this. The blue section represents the profit center, the members who drive our earnings. With the advent of AI, I believe that junior-level employees with annual incomes below JPY 6 million will become less necessary. In that sense, we quickly stopped hiring members with salaries below JPY 6 million. Also, as mentioned earlier regarding AI application for corporate teams, we've stopped hiring 95 people there too. Over this second half, as I said, we plan to end contracts for all 110 temporary staff as they expire. If we can achieve these goals in this second half, then even as we aim for JPY 190 billion in sales next year, I believe our workforce structure for the JPY 300 billion target in fiscal year 2030 will shift like this. The key point will be whether we can reskill our current members into consultants and engineers who can effectively provide services in an AI-integrated environment. Regarding that, we've already created skill assessments. By reskilling, we've identified 4,248 members with the aptitude to actually do this, so I feel we're well prepared for an AI-integrated future. Also, for our back-office members, we plan to move about 100 of them to the BPaaS unit by next year, where we want them to thrive as members who generate revenue in a profit center. When sales reach JPY 300 billion, reskilling for AI native will become extremely important. For that, we've already developed the skill assessments. We're currently in the process of transitioning our reskilling efforts, and we intend to prepare thoroughly for that JPY 300 billion sales target. At that stage, we plan to operate with our current back-office staff without any new hiring, maintaining our existing headcount. Based on our internal estimates, we believe we can achieve the gross profit, SG&A ratio, and operating margin shown here, so we're committed to working toward those goals. In summary, it's been a whirlwind year for SHIFT regarding AI. Despite the constant stream of news, we've successfully integrated current AI technologies and leveraged them to generate actual sales. We've built services that drive revenue and established a framework that effectively boosts our gross profit. Through these efforts, we aim to build a solid path to JPY 200 billion and JPY 300 billion in sales, becoming a leading Japanese company that truly utilizes AI. We would appreciate your continued support for the SHIFT Group. Thank you very much for watching until the end today.