ExaWizards Inc. (4259) Earnings Call Transcript & Summary

Good evening from Tokyo Japan, and good morning to most of the attendees here. Welcome to ExaWizards' earnings call for the fiscal year ended March 31, 2023. This is Koichi Kawai, Head of President Office and Deputy Head of Corporate Management of ExaWizards. With me on the call today are Makoto Haruta, Representative Director and President; Tomoaki Maekawa, Head of Corporate Management; and Hikaru Koyama, Head of Investor Relations. After the market closed today in Japan time, we issued a press release with details regarding our fourth quarter results as well as supplemental data sheets and FAQs to our results, both of which can be accessed on the IR section of our official website. Please note that this meeting is also being recorded and will be posted on our IR website at a later date. During today's call, we will make statements relating to our business that may be considered forward-looking. These statements reflects our views only as of today and should not be considered representative of our views as in any subsequent date. We disclaim any obligation to update any forward-looking statements or outlook. With that, I would like to introduce today's schedule. Today, Mr. Haruta and I will begin by explaining the financial results, growth strategy and earning forecast for about 30 minutes followed by the presentation materials. After the presentation, we will have a Q&A session until 11 p.m. Japan time. Finally, at times in our prepared comments and response to your questions, we may discuss metrics that are incremental to our usual presentation to give greater insights into dynamics of our business or our quarterly results. Please be advised that we may or may not continue to provide this additional detail in the future. With that, let me turn the call over to Mr. Haruta, our Founder and President, for his prepared remarks. Mr. Haruta?

Good morning to U.S. and Europe and good evening to Japan and Asia attendees. I'm very excited to meet you, investors and analysts, globally. This is Makoto Haruta, the Founder and the representative director of ExaWizards. Today, we will provide you with the earnings results on the fiscal year ended March 2023 as well as our strategy for this fiscal year which I'm very excited to share with you. I have started this business in 2016. And now in the most exciting environment where since I have envisioned in the past, is actually started to Japan -- to happen. For instance, under the environment where generative AI became one of the big theme globally, we are seeing enormous demand and traction from a huge number of enterprise clients across Japan, which is expected to boost our AI Platform business. And here, solving social issue of productivity of Japanese enterprise is a huge issue we are facing and seeing huge demand. Also, our flywheel service model to sort social issues through artificial intelligence is seeing enormous traction in DX Assessment & Learning for enterprise and Care service under the most aging society in the world which is expected to boost our AI Products business. Just to add, we have announced several product already with ChatGPT which we have seen strong traction from large enterprises and white collars in Japan. I'm very excited in the current environment and looking forward to this new fiscal year. Today, Kawai-san will explain the details of the presentation and followed by his presentation. I'm open to answer any questions from you. Thank you very much.

Thank you very much, Mr. Haruta. And please move over to Page 4 for our executive summary for the fiscal year ended March '23. So sales for the fiscal year totaled JPY 5,591 million with Q4 sales of JPY 1,583 million. The growth year-on-year was 16.2% and growth for the quarter-on-quarter -- year-on-year was 3.9%. We apologize in advance for any concern to the revised earning forecast may have caused. But although it was not enough to add to the numbers, we did achieve the revised planned profit and our expectation here was actually way higher in the early year. And in addition, the extraordinary income was announced as JPY 293 million, the one in the first quarter and a gain on the sale of Hanasuto was recorded in the fourth quarter. And moving on to the segment information. The AI Platform segment, the revenue was JPY 4.7 billion, up about 12.5% year-on-year. Despite a drop off of the large clients, the overall number of clients was 142, the increase of approximately 20% and top 10 clients revenue per client was JPY 229 million, a positive year-on-year results. Although we had a loss of several large clients, we were successful in [ valuating ] the growing of ARPU. And the long-term clients, which we define as revenue from clients who's paying us for 4 consecutive quarters and more accounted for 73.9% in the total. By simultaneously growing these clients who has a long-standing relationship with us, and we will provide -- continue to increase the number of new clients as well. And moving on to the AI Products business. The AI Products business totaled JPY 888 million, increase of 40% year-on-year growth. Including the Hanasuto sales I mentioned earlier, the business is growing steadily and operating loss has narrowed down to JPY 45 million, which is gradually becoming close to positive. For the fiscal year ending March 31, 2024, we are forecasting JPY 7 billion for the full year with approximate growth rate of 25% and operating income of negative JPY 370 million. We plan to increase sales, although operating income will be at the same level of this fiscal year. As seeing the current environment mentioned by Mr. Haruta of booming AI demand, we want to make aggressive business investment and promote development of new and existing products while utilizing those generative AI well. Now let me go over a high-level overview over the March '23 by skipping to the Page 7 on the consolidated financials. So the sales as shown in the graph, and operating income is below the last year's level. And Page 8, please. This is a quarterly trend for the past 3-year consolidated basis. Usually, we tend to grow strongly over the first quarter to the fourth quarter. However, the impact of the sluggish sales growth due to the loss of several large accounts in the third quarter was significant. And despite our efforts to recover in the fourth quarter, the recovery was limited to a certain extent. However, we are not [ pessimistic ] to the next year, and we're very much excited to see the growth in this segment with many traction from new clients now. And moving over to Page 14 on the specific AI Platform segment. Overall, the sales for this fiscal year ending March 2023, were JPY 4.7 billion, increase of 12.5% from previous year. On the other hand, regarding the profit, the increase in headcount for business expansion has resulted in operating loss as the increased costs could not be covered by sufficient sales. Now moving over to Page 15 for the quarterly results. On a quarterly basis, the number of employees, including new graduates, tend to increase at the beginning of a new fiscal year, while increasing number of employees we train and develop them to make a good revenue and results but for this fiscal year, the growth in the third quarter and fourth quarter figure was below expectation, resulting in the weaker operating profit situation in total, as I mentioned earlier. However, we are currently making steady progress in acquiring new clients and in particular, inquiries about generative AI is strongly increasing and we assess that the business environment is recovering strongly. And I will explain these points later and move over to Page 16. Even with the last of the few big clients, the number are maintained well with average sales of the top 10 remaining at the same level over the previous -- last year. Meanwhile, the right-hand side, that number of clients grew steadily. So overall, we do not assess the business environment as deteriorating and believe the company is in a state of good and healthy strength. And Page 17, please. As we've seen the sales from the long-term clients to grow healthy, we expect to show strong results on the efforts of making this as big accounts while becoming new enterprise accounts as well improved in the performance, and we have a strong confidence to do so this year. Now moving over to the AI Products segment. Page 19, please. The progress of the AI Products segment has been growing steadily. Sales growth have been JPY 888 million, which is 40% year-on-year growth. And for your reference, this also includes the sales of the service that we were partially transferred as explained earlier and moving on to 20. So this is a quarterly revenue trend for the past 3 years, and the sales has been growing significantly despite adjustment of the divested businesses. As can be seen in the graph of the quarterly results, sales trend [ further ] increased each quarter in a healthy manner. Also to mention on the revenue for this fourth quarter, that CareWiz Hanasuto service contributed in the part, and the figure was JPY 340 million and 64.7% year-on-year growth. Looking ahead to the next fiscal year, CareWiz Hanasuto is now being transferred out of the business, so this portion will drop off from the next fiscal year. And moving over to Page 21. With annual revenue with service segment, the bottom blue parts, DX Assessment & Learning service in the bottom is growing significantly. And Hanasuto, which is in the middle part, the social AI [ Care ] product category. But even if this product is transferred, the next new products are there to be introduced so that in total, we will finally increase overall product sales for the next year. Now moving over to the growth strategy with Page 23 and beyond. So this is one of the most important pages and I would like to discuss how we intend to grow our business performance in the current fiscal year. This chart summarizes our current environment. The company was originally founded in 2016 by Mr. Haruta, but AI has been evolving rapidly since before then. In 2012, Professor Hinton won a competition for image recognition. And from that point on, image recognition in AI came into the limelight. In Japan, Professor Matsuo of University of Tokyo and others publicized book in 2015, which made people more aware of AI, what deep learning is and the potential of AI. And after that, there was a breakthrough in 2016 to '17 when AlphaGo beat be a Go player. In the midst of all this, we founded the company in 2016. After that, there were times when people thought AI to be great. And other time, people thought it wasn't so great. However, the deep learning has been evolving rapidly since around 2013 and Transformer and GPT-2 were created and what is called generative AI has been evolving. The generation of image in 2021 and '23 was only the part of this area, and there were a lot of excitement about being able to create images and pictures. Then at the end of 2022, ChatGPT was able to generate text. Those who are non-engineers or non-designers tried out ChatGPT, they became aware of the great potential with generative AI, and there were renewed interest in AI. As a result, GPT evolved and became used by more people and attracted more attention by the end of year 2022. While thinking about what AI can do was original theme of the company, the stage has clearly changed in us in terms of creating something with AI. The number of employees and number of customers is increasing significantly since the company was founded in 2016. However, even excluding this, our customer feedbacks, reactions and awareness and enthusiasms of engineers, including our business employees are clearly higher now. And we're very much excited with this trend. Moving over to Page 24. Regarding the generative AI, we're seeing enormous contact and demand from enterprise customers on how to utilize them in Japan. In the seminars that we have been holding recently, those with the theme was generative AI has attracted many more visitors ever as you see on the page, left-hand side. And with Mr. Oue, the member of the Board of Directors took the stage at Japan IT Week, many people attended, more than over 1,000, and it took more than 30 minutes just to exchange business card with all the line after his presentation. We've seen a huge demand and possibilities of how we can meet the hot demand of our customers. Even in this kind of environment, the company has not changed management access. And we believe that we're now in the business environment where we can finally give shape to what we are doing. And 25, please. To reexplain this diagram, what we want to say in the last -- is that left circle is that we originally founded our company while we did not know how to apply the new AI and machine learning as a technology to be the most effective for our customers. Rather than recommending the best way, we started by exploring the use cases together with our customers on the left-hand side of the AI platform. As a result of solving individual customers, clients' issues the new AI models' data, algorithms have been accumulated. In the process of solving those individual problems, we took a step back and recognized that there are so many problems and issues in the industry and society as a whole, and we proceeded to expand the business for the general social problems. We started from the point of sublimating these issues into generic issues, then providing solutions to the generic issues as a services. When generalized issues are offered as a service and used by more people, costs will go down and many people's issues have been solved. And this is a cycle on the right-hand side, the AI products business. If we apply to the individual companies and new issues arise, we will again enter the cycle to the right to the left. While the left side is growing steadily, and the right side has not yet reached to its full potential yet. But with our products such as DX Assessment & Learning and Care businesses are building up as a full step on the left-hand side. Through the repeated efforts, our employees are working hard and we have nearly 100 patents, which is #1 in public AI start-ups, which is a result of solving large enterprise problems through our AI Platform business in the past. We are now ready to develop our business to take advantage of this and very excited with this opportunity. And Page 26, please. Mentioned on our business model in detail. The major theme that has emerged in the course of dealing with clients was the need for companies to improve their own productivity in the face of a social need or improving corporate productivity. The 3 years with COVID-19 has intensified this awareness in Japan as well. The first thing the company did was the #2, that improved productivity on the site. With aging society and the workforce in Japan, it is necessary to secure the on-site workforce. And we also need to pass on the technologies, but there are no recipients and workforce is decreasing here. To solve this, as exaBase Robotics, we have been working intensively to improve productivity in the field of multimodal AI. In this way, we found that even in a corporate system development, it is quite difficult to link the new services of AI with old business process even though the company has a desire to create new system utilizing AI. And by learning that regarding #3, we are working to streamline those AI development. For example, there is huge issues in Japanese corporate with a limited skill set of the internal engineers so that they have to rely on the external parties even on small software improvement, not knowing what's going on with a huge budget. We have been developing exaBase Studio to solve this issue since last year to make this way way easier. And we are working on the theme of including productivity by improving the efficiency of the system development. And about the first issues, as I mentioned earlier, exaBase Assessment & Learning is a service that address the development of enterprise customer skill set. There are huge issues in Japan of not knowing where to find the right internal people for such AI or DX efforts and how many skilled employees are located in which departments, although it's important to develop digital human resources. This may sound funny to global investors, but this is a real issue in Japan, so the government is pushing the company to reinvest more for the assessment as well as education for the corporate employees for the DX. And about the first one on the top, our original goal was to improve productivity in the field but the biggest issue was the productivity of the white collar workers. And when we thought about what to do about it, we came up with just generative AI. By utilizing AI, we believe that we can increase what white collar workers can do as a whole and getting customers to work on it, we can increase white collar productivity. We're promoting our business for improving corporate productivity based on these 4 axes. With regards to some operational changes, please take a look at Page 27. First of all, we have made ChatGPT available to all the employees. And in addition to that, quickly set up an operational reform group directly under President Haruta and also created a system and team for engineers to support this from a technical standpoint. And Page 28. Also, the products that have been created as a result of this -- our exaBase Generative AI and exaBase IR Assistant. First, let's talk about exaBase Generative AI. When you want your employees to use ChatGPT and especially in Japan, there are huge issues such as complexity of having each employee pay individually was security issues such as possibility of confidential company information being uploaded or not knowing who is utilizing ChatGPT and how they are using it. There was a need to understand ChatGPT for the perspective of overall management in the Japanese enterprises. These services were created as a result of examining what the minimum requirement from those Japanese corporates' standpoint. We have now released these services in the hope that a number of its users will increase if we can provide secure environment at the same time providing its original prices. And we're seeing huge traction in these products already. In addition, I have -- we have another exaBase IR assistant that we have just announced yesterday. In Japan, during the earnings results or shareholders meeting, many enterprise create 300 pages of Q&A to prepare for the investor questions. And this IR assistant was created to lighten the burden on such people as much as possible. It is a service that can create both questions automatically and answers as well by reading the various financial documents. And this alone is not enough to ensure accuracy, the system is designed to improve accuracy of questions and answers by having a human participate providing the feedback as well. We have just started expecting -- we have just started accepting the reservation for the beta version, and we're seeing many traction from clients already. And this was a huge news on the Nikkei newspaper article as well. 29, please? And second, as I mentioned earlier, exaBase Robotics for the multimodal and productivity improvement, we're continuously working with several clients, including Nippon Steel Corporation and JAXA as shown on the right-hand side, to directly address their on-site issues. In the midst of this situation, we have now come up with the idea of generative AI. We recognize this as a great opportunity. And we believe that combining the technologies we have seen and cultivated in the past with the new AI technology will lead the new evolution that in the future, robotics will be able to controlled by the speech and other means, which will further promote their widespread use cases. We believe that this area where exaBase Robotics and large language model will be developed in the tandem manner. And page 30, please. Regarding the system development, exaBase Studio as explained in the past, and we're thinking of developing the system that can be used by non-engineers that can be easily written in the GUI and that can be used by engineers and non-engineers so that it is easy to understand what we're doing as shown in the image on the right-hand side. This will enable us to develop a system that we can operate on our own. And in other words, without the need of outsiders. And now that we have heard about the generative AI, we believe that the use of generative AI will enable engineers to evolve their own services using such a system. And Page 31, please. And at present, about 60,000 people from 900 enterprise around Japan has been taking this DX Assessment & Learning test. As a result, we have data on the level of each of the various companies in the various industries. This allows us to understand the relationship with our clients in order to estimate the level of their own company. After determining their level, we will recommend appropriate content for the growth and development and if clients implement a training program based on the content, the skills of the employees will improve as a result. And we have the system that goes around a circle in the form of further assessment and this and that. If we can spread the system to the rest of the world, I believe that we will be able to surely increase the level of digital human resources. And we're currently updating the program so that it will be compliant with the DX literacy standard, the Japanese government, METI, the Ministry of Economy, Trade and Industry is providing. And as a result, we would like to develop this product in line with the worst trend in the future. And so due to the sake of time, please move over to Page 34. If we have to do everything by ourselves to introduce new system, it will be not that efficient, and we're undergoing the major restructure reforms. The first major pillar on the left-hand side is a review of the business portfolio and the second is the way business is operated should be conducted. On the left-hand side, regarding the business portfolio, we have transferred some services in the area of Care in the first quarter, and we also transferred Hanasuto the fourth quarter of this fiscal year. In addition, we had an Indian subsidiary as a development and research basis. And today, we announced the decision to close the Indian subsidiary as a structural reform. Although this impact has already been factored into our figures for the estimate, we are looking to improve our portfolio in this way, while commonly looking at the timing as some needs at this time, what we should be doing. On the other hand, on the right-hand side, we have increased the number of people as our service and company grow in order to manage them well but we believe that we must always aim for lean structure, not one in which everything is solved by people. The cycle on the left-hand side, is a diagram that shows individual company model -- let me move over to the next page for the earnings forecast. Page 36 with the full year earnings forecast for the fiscal year ending March 31, 2024, with a tremendous expansion of what we can do. We want to manage the company by keeping a close eye on the numbers as we know that if we do not control our own cost, we will lose our footing as a company. As a result, for the fiscal year ending March 31, 2024, as mentioned earlier, we expect sales to be JPY 7 billion and profit level on the similar with the fiscal year ended last year. And the important point is mentioned in Page 37. A bit on the final [ point of ] audit, the existing business are profitable as they have expenses just over JPY 6.8 billion for the sales of the JPY 7.0 billion. In addition to this, we are also considering investment in area where we have not yet forecasted sales. And where we hope to increase sales further by making investment in additional to the existing services. Our investment is mainly in the personnel expenses so we will secure human resources and talented employees and allocate them within the company to the investment. We will generate profit from existing businesses while at the same time, creating new businesses as well. As shown on the right side of the chart, our major challenges for this fiscal year is to make investments that will lead to the new sales as shown on the reference over the chart. Next page please. Regarding investment of the fund and shareholder returns, as we continue to make solid progress in our business performance, we, as a company, must meet expectations of our shareholders and investors. And once again, as I have stated, the first and foremost thing we must do is to raise the corporate value, the enterprise value. There may be many ways to raise the stock price. And one is, of course, to raise business performance. However, before that can be done, the company needs to raise our own expectation. And when we look at the stock price after the listing, we must first rate it to the original price where investors can actually believe into us. In order to achieve this, we must actively develop the variety of new products. And to this end, we will aggressively develop various products and expand existing services while firmly executing the measures and producing more and more results which we believe will lead the business performance. In the process of building a company that can meet everyone's expectation, we want to raise the share price. As a part of this process, we are acquiring treasury stock as a part of our financial strategy. As announced recently, we intend to use this treasury stock for potential M&A and potential other purposes in the future. In addition to expanding the performance of our own services, we would also like to expand the company through our financial strategy as well. We're thinking in this way. So thank you very much for listening. And let me pause here for potential questions from the investors. Thank you very much.

[Operator Instructions] We have a comment from a [ Robocaps Tech ], nice meeting you and let me read out some of his questions for us to answer back. So how could generative AI contribute to the revenue in the fiscal year March '24? How much additional revenue would be generated? Can I please pause and let me discuss this Mr. Haruta before getting back to you with answers. By interpreting what Mr. Haruta has mentioned, at the current environment, we're not expecting the revenue. We're not incorporating to our forecast the potential new revenue that would be generated from generative AI services. And answering back to your second question, besides introducing GPT-4, does ExaWizards have any in-house generative AI know-how? Let me pause before getting back to you.

[Foreign Language]