$30 off During Our Annual Pro Sale. View Details »

Overview of ABEJA, Inc.

ABEJA
December 16, 2021

Overview of ABEJA, Inc.

ABEJA

December 16, 2021
Tweet

More Decks by ABEJA

Other Decks in Business

Transcript

  1. Implement a Fruitful World 1

  2. 1. Corporate Profile 2. Business Overview 3. Work Environment, Team

    and Culture 2 Confidential 2
  3. 1. Corporate Profile 3 Confidential 3

  4. Confidential 4 Company profile Name of the company Head office

    Foundation The number of employees Board members Technical advisors Representative Director, CEO Yousuke Okada Representative Director, COO Motohiro Koma Director, CFO Kazuki Hanabusa Outside Director  Kunihiro Tanaka Outside Director  Koji Asano Outside Director  Ryosuke Takatsuki Outside Director  Hitoshi Matono Full-time Auditor  Junichi Tobaru Outside Auditor Takuma Shimizu Outside Auditor    Masaaki Aoyama Hitoshi Matsubara Professor at Graduate School of Information Science and Technology, the University of Tokyo Yutaka Matsuo Professor at Graduate School of Engineering, the University of Tokyo President of Japan Deep Learning Association (JDLA) Confidential 4 ABEJA, Inc. the ARGYLE Aoyama WeWork 6F 2-14-4 Kita-aoyama, Minato-ku, Tokyo 107-0061, Japan September 10, 2012 95 (as of March 16, 2022)
  5. Confidential 5 ABEJA’s philosophy Tagline: Implement a Fruitful World We

    drive business with our philosophy “Implement a Fruitful World”. We leverage our own strength - “collaboration between AI and humans” and “fusion of digital and real”. We lead our client companies to become indispensable existence in the future world. Confidential 5 Value: Technopreneurship Move Fast Focus on Impact Update Yourself Team First Win Right Mision: Reform industrial structures by the power of technology Vision: Change the world through our innovation
  6. Confidential 6 Our code of conduct Entrepreneurship Technopreneurship Liberal Arts

    The driving force to promote circulation of the both Keep questioning ourselves our own conducts by Liberal Arts Innovate by technology and make impactful contributions to society Society Culture Driven by entrepreneurship, we continue to question ourselves how society should be from both liberal arts and technology points of view. At ABEJA, we call those who embody such behavior “technopreneurs”. In today’s world, which is undergoing major changes, we believe that technopreneurs are the ones who will create new values with breakthrough ideas. They are the ones to bring about a truly fruitful world. Technology Confidential 6
  7. 2. Business Overview 7 Confidential Confidential 7

  8. Confidential Our Business We provide professional services to our client

    companies to meet their individual needs for DX. Also, we provide versatile systems and services to multiple companies to solve common customer problems. Our product, “ABEJA Platform”, enables efficient AI deployment and it lets us provide services that exceed customer and social expectations. Highly common and reproducible problems Speed up business and better efficiency Highly common and reproducible problems Speed up business, better efficiency collaboration Mutual *Including not only fully AI-automated systems but also the systems that were designed to work with “Human in the Loop/ Expert in the Loop” for the actual operation in live environment. Solutions We provide professional services to client companies to meet their needs for DX individually, by “ABEJA Platform”. Upper process Grand design, DX/AI strategies Middle process BPR, Model dev, evaluation, operation, HR strategy Lower process Integration, , BPO, DX/AI trainings Operation process Live operation (HITL/EITL*), retraining, growth Products We build various systems on “ABEJA Platform” as versatile systems and services, and provide them to multiple companies. Horizontal Recommendation system Repair parts identification system Demand prediction system Vertical Retail Distribution Real estate etc. etc. It enables AI incorporation to propel DX processes throughout. The software suite consists of various functions that are well organized and integrated. Our tech partners Confidential 8
  9. 9 Our solutions reach production and receive excellent customer evaluation

    We have excellent technical capability such as image analysis, natural language processing, etc., leveraging advanced Machine Learning technology. A lot of our solutions have reached production, and achieved DX of over 300 client companies. Highly evaluated by industry leaders • We are the first company in Japan to receive investment by the US headquarters of world leading companies such as Google and NVIDIA. • Various prestigious Japanese companies also invest in ABEJA. • AWS evaluated our technical capability. We are the first company in Japan to be certified with Machine Learning Competency. Our investors Our solutions have been deployed to over 300 client companies • Since the start of the AI boom in 2012, we’ve deployed various technologies, particularly Deep Learning, to various industries during the past 9 years. • Our solutions have been introduced and deployed to over 300 companies in various industries (retail, manufacturing, logistics, infrastructure, finance, medical, etc.). We’ve been gaining a lot of experience from operation in the actual business environments. Manufacturing Logistics Retail Finance Infrastructure Medical IT, Telecom, Media Confidential Our clients Confidential 9
  10. Confidential Solutions 10 Confidential 10

  11. 11 Our client companies Business partners Provide systems Provide services

    Pay Highly common and reproducible problems Speed up business, improved efficiency collaboration Mutual Solution field We provide professional services to client companies to meet their needs for DX individually, leveraging “ABEJA Platform”. Upper process Grand design, DX/AI strategies Middle process BPR, Model dev, evaluation, operation, HR strategy Lower process Integration, , BPO, DX/AI trainings Operation process Live operation (HITL/EITL*), retraining, growth Fee Product field Provide services Fee 1 2 3 4 We have advanced technological capabilities and wealth of knowledge backed up by experience. We provide solutions that drive successful DX of our client companies. We tackle core problems as a partner of leading companies in various industries. We strive to reform industrial structure and business models. We go deep in customers’ business, gain a lot of domain knowledge and identify common customer problems. We develop and provide systems to solve them. We make DX happen with speed by “ABEJA Platform”. We gain new knowledge and experience through various projects, and they are feedbacked to “ABEJA Platform” which will further speed up DX and enhance the impact. 1 Features of our services 2 3 4 Confidential 11
  12. Success stories - Joint business with Hulic ABEJA and Hulic

    have been developing various systems for office DX since November 2020. Hulic highly evaluated the accomplishments we’ve made from the joint business in the last one year as well as ABEJA’s human resources, technology, and experiences in AI/DX in various industries. As a result, we have concluded Capital and Business Alliance Agreement in October 2021.
 
 Hulic is a leading company in the real estate industry. We will further strengthen our partnership and drive the joint business. We will reform workplace and contribute to solve social problems through our business in the real estate industry. By these activities, ABEJA aims to achieve our vision “Implement a Fruitful World”.
 Confidential 12 Main initiatives and achievements We’re driving development for “Bizflex by HULIC”, a professional rental office solution that combines essential elements for today’s workplace. On November 1, 2021, HULIC opened a medium-sized flexible office building, “Bizflex Azabujyuban by HULIC” that embraces three concepts - Subscription, Workplace, Office. Confidential 12
  13. Success stories - Joint business with SOMPO Holdings ABEJA and

    SOMPO Holdings started joint business in 2020. To handle insurance business in Japan, we’ve worked on data analysis, built prediction models utilizing Machine Learning, carried out joint business development, and developed DX human resources and the hiring systems.
 SOMPO highly evaluated the achievements we made from our joint business as well as ABEJA’s technical capabilities, products including SaaS and Paas, and human resources including our engineers and Data scientists. As a result, we have concluded Capital and Business Alliance Agreement in April 2021.
 
 Confidential 13 ABEJA will contribute to promptly achieve SOMPO Holdings’ objective, “Real Data Platform for Security, Health, and Wellbeing”, making the most of real data created by the SOMPO group. Through these activities, ABEJA aims to achieve our vision “Implement a Fruitful World”. Confidential Existing areas New areas Quickly realize Real Data PF for Security, Health, and Wellbeing Promote AI utilization Develop digital human resources Combine the data SOMPO group accumulates and ABEJA’s tech expertise to promote AI utilization in their core biz operation (e.g. Insurance underwriting, etc.). Make various systems to hire/retain DX specialists and base human resources based on ABEJA’s 8 years of experience. Cultivate the culture, etc. Leverage SOMPO & Palantir’s strengths and ABEJA’s expertise to drive overall biz dev. Start with product development to meet needs on site in elderly care business. Not only incorporating AI or achieving DX of conventional App layer, but also provide added values to digitize core system, considering data value and utilization Make systems to achieve true “in-house” DX human resources hiring and nurturing
  14. Motion analysis of factory workers (Denso International) Problem They used

    to measure the time spent for each process manually to improve factory workers’ operation, which they wanted to automate. Solution We used Bone estimation technology and built an AI model that can automatically measure the time spent for each process using cameras. Reference media article ABEJA supported kaizen of a factory for Denso Thailand Detection for risk mitigation (Hitachi Logistics) Problem The client wanted to find risky driving using dash cam info, and give instructions to the drivers to avoid traffic accidents. But it was hard to find high-risk driving from big video data. Solution We made dangerous driving videos into training data, let an AI model train with the dataset, and built a model that detects risky driving automatically. Reference media article Save people’s lives by AI! Hitachi explains how they made “Safe driving management system” Automated eye inspection (Topcon) Problem To further advance Ocular fundus examination, the client wanted to utilize AI and automate the examination. Solution We created training data with medical doctors, established an AI model for automatic ocular fundus examination to automatically examine fundus and specify the pathological point. Reference media article Presented at SIX: AI utilization for ophthalmology images - What AI can bring to clinical practice Other projects, examples (1/3) Confidential 14 Confidential
  15. Confidential Other projects, examples (2/3) 15 Industry Theme Project overview

    (objects and detection targets) Types of data used Any Marketing recommendation Customer clustering and recommendation engine development Customer data, purchase data Any Knowledge search Advanced knowledge search for customer helpdesk and internal operation Existing knowledge data, CRM data (communication history) Any Text Analysis Judge the severity level of customer complaints received from consumers to the customer helpdesk Past history data of complaints Any Copyright search Judgement of whether if the products to be purchased and displayed violate copyright or not. Product image data, description Any AI-OCR Achieving output with 100% accuracy by using HI (BPO) to assure AI accuracy Scanned image data of forms Manufacturing Behavioral Analysis Factory workers’ motion measurement model for a manufacturing factory Video data of experienced workers’ work movement Manufacturing Visual inspection Visual inspection of car parts Image data of the target areas of the parts to be visually inspected Manufacturing Demand prediction Demand prediction of product sales Past purchase data, customer data, factor data that may correlate with demand Manufacturing Establish a simple model for Advertisement copy review Detect faulty expressions, wrong product names or approved names in Advertisement copy texts. Faulty expressions list data, product and approval names master data Manufacturing Maintenance support efficiency Predict repair parts to be brought to the repair site based on the inquiry contents. The last few years’ inquiries and communication history data Infrastructure Infrastructure inspection Inspection of rust/corrosion on bridges using drones Image data including the target construction/building Infrastructure Facility inspection Detect the degree of corrosion on factory buildings Image data including the target construction/building We work as a partner for clients in various industries conducting various types of business. We support client companies’ business reform by DX and we challenge business innovation by collaboration between AI and humans, in order to promote industrial structure reform. Confidential
  16. Confidential Other projects, examples (3/3) 16 Industry Theme Project overview

    (objects and detection targets) Types of data used Infrastructure Railway inspection Inspection of train railway Image data including the target construction Infrastructure Scrap copper estimation Estimation of scrap copper (breaking down to subdivisions) Image data per type Infrastructure Estimation of danger levels Objects/positions where incidents may happen (building exits, etc.) Image data, sensor data Logistics Estimation of anomaly Detect anomaly of logistics equipment Sensor data and anomaly data of the target object to be detected Logistics Visualize packaging Measurement of the time spent for packaging and the number of packed items Video of packing work of the target products, image of the target items to be packaged, classification linked to the items to be packaged Logistics Prevent dangerous driving Raise an alert for dangerous driving based on multiple definitions Data of drivers obtained by sensors Construction, Logistics Risk/incident mitigation Raise an alert when people come too close to machines Video data of the target in the factory to be detected Game Cheat detection Detect unauthorised accounts from action log in a game. Auto monitoring of field activity log, etc., by AI to check/monitor accounts conducting unauthorised actions. Action log in a game, battle history, field activity log SNS Detect unauthorized posts (social media function) Auto monitoring by AI to check if posted texts, images, videos, etc., by users include inappropriate contents (abusive language, offensive to public order and morals) Image, video, texts Media Advertisement CTR prediction Use past Advertisement distribution data to train image features and actual results to predict the degree of effectiveness by the features of new images Image Advertisement data (image file, number of display, number of clicks, number of conversion) Media 1to1 recommendation Optimize display of contents in the client’s service User data (purchase log, user master data, feedback data), Metadata Communicatio n Demand prediction for store sales Demand prediction for the number of store sales per store/ per product Past purchase data per store, customer data per store, Factor data that may correlate with demand per store Leasing Total platform Maximize profitability by dynamic pricing Past sales history data, stock status data, external input data (Big data - events, weather, competitors, social media, etc.) Medical Automated medical inspection Automated ocular fundus examination by utilizing AI Medical image data Confidential 16
  17. Confidential 17 Products (ABEJA Insight for Retail) Confidential 17

  18. Acquire data POS data ToF sensor Camera Integrate, analysis Visualize

    Image analysis Dashboard Hypothesis ? Analysis, Verification Measures • To make right decisions on measures to enhance sales, and grow out of having too much dependency on individual staff’s experience and intuitions. • To explain the reasons why the sales has improved • To expand application of effective sales enhancement measures to other stores Service overview “ABEJA Insight for Retail” is a service for retail stores. It visualizes and analyzes customer behavior to enhance reproducibility of “the state of selling well”. ABEJA Insight for Retail enables: Confidential Confidential 18 18
  19. Confidential Functions overview We introduce the right equipment for measurement

    according to the purpose, based on the problems in stores and between store managers and head quarter Target setting Measure in-store customer behavior by utilizing cameras and sensors ? Once measurements are done in stores, the information is organized appropriately for each role - store managers, area managers, head quarters, etc. Stores and the head quarters can verify the effectiveness of the implemented measures, then identify the next action to take, supported by our Customer Success team. In-store measurements Dashboard, Report Verification of the effectiveness Confidential 19 Our focus is not just providing data but “letting our clients utilize the data”. Our service covers “target setting” and “verification of the effectiveness of measures” as well as measurements and reports. It supports clients with building natural habit of data utilization in their daily work. 19
  20. Confidential What can be visualized by in-store measurement 20 Cameras

    installed in stores measure storefront traffic and visitor counts real-time Obtain facial image and estimate age/gender by image analysis Measure how many customers came in and stayed for how many seconds. Understand customer flow. Estimate new/repeat customers. Measure indexes such as “x% visit the store once a week/month””, etc. Storefront traffic, store visitor count Age/gender estimation In-store customer flow Repeater analysis We make the most of IoT × AI (mainly image analysis) × BigData × Web technology, which can visualize the following information. Confidential 20
  21. Confidential ABEJA Platform 21 Confidential 21

  22. Confidential ”ABEJA Platform” is our own service. It enables “collaboration

    between AI and humans” and “fusion of digital and real” which are essential for DX. Through our service, we build platforms together with our clients and let the solutions reach production and move onto a continuous operation phase. It’ll lead to reform the clients’ core business, and we challenge to innovate business models.
 API Retail Manufacture Infrastructure Construction Logistics Cloud On premise Data center Supports the entire DX processes Edge 22 Service overview e.g.) Establishment of infrastructure such as data center, storage, GPU. Model revision management, O&M systems, etc. Confidential 22
  23. 3. Work Environment, Team and Culture 23 Confidential Confidential 23

  24. Our organization 24 ABEJA provides professional services to meet the

    needs for DX of client companies individually. Therefore, we have a lot of members and data scientists in the Solutions business field. Confidential Confidential 24 Composition by business field (Excluding the ones with multiple roles) Products CEO Office AI Product Div Solutions Composition by job type Corporate Engineer & DS Division, Business Corporate Business
  25. The way of working at ABEJA 25 Remote work and

    Flex time • To let people work flexibly, we have Remote Work and Flex Time systems. • The working hours are 8 hours per day including the core time 10:00-15:00. Free to use WeWork offices across Japan and “Bizflex Azabujyuban by HULIC” • We have exclusive office space at WeWork in Gaienmae as ABEJA’s head office. • We also have ABEJA’s exclusive office space at “Bizflex Azabujyuban by HULIC” to explore office DX. • In addition, our employees can freely use WeWork offices across Japan. We support flexible and diverse ways of working. Confidential Confidential 25
  26. The way of working at ABEJA 26 A special leaves

    to recharge your battery • We have employees from various countries, so we don’t have a fixed summer holiday. Leaves are granted in the form of additional paid vacations. ◦ Those who joined the company during 2nd April till the end of June: 3 days ◦ during 1st July till the end of September: 2 days ◦ during 1st October till the end of December: 1 day ◦ during 1st january till the end of March: None Year-end and new-year holidays • Four days of designated paid holidays are granted for the year-end and new year holidays. Annual paid holiday • Upon joining the company, employees are granted 10 days of paid vacation. Then, it will be granted based on years of service. • Employees can use paid vacations from the probationary period. Years of service Upon joining 1 year 2 years 3 years 4 years 5 years 6+ years Days granted 10 days 11 days 12 days 14 days 16 days 18 days 20 days Confidential Confidential 26
  27. Reference: ABEJA’s response to COVID-19 27 ABEJA continues to be

    flexible, giving top priority to creating a safe and secure work environment for our employees. Confidential Confidential 27 Employees are allowed to work remotely completely • Before the outbreak of the new coronavirus infection, employees were allowed to work remotely up to 2 days/week. However, currently we allow employees to work fully remotely. (*1) • We have some employees who live and work remotely in Nagoya, Kyoto, etc. Paid Leave for COVID-19 Vaccinations • We have special treatment on work hours regarding COVID-19 vaccinations: ◦ Who is eligible?: Full-time regular & contract employees, Part-time employees who are covered by corporate social insurance ◦ If an employee gets vaccination during regular work hours on a weekday, the time required for vaccination including transportation, is regarded as worked hours. ◦ The time required to recover from any after-effects from vaccination within 3 days is also regarded as worked hours. ◦ If an employee takes a full-day break using a paid holiday, we provide another paid holiday to make up for it. *1: We have not yet decided the policy after the outbreak of new coronavirus infection has come to an end..
  28. 28 Employee benefits and allowances (examples) Challenge system • The

    company provides resources (cost, location, etc.) for employees to take on a unique challenge that is not limited to their current work or mission. (They need to apply the project plan in advance.) • Employees are free to work on the challenge outside of their regular work hours. • Examples ◦ PoC of face recognition payment system at the office kiosk ◦ Application for patents ◦ Making an AI×IoT workshop package ◦ Get qualifications  ◦ etc.               Childcare allowance • We provide 10,000 yen per month for each child up to 20 years of age for whom you have parental rights. (Starts after the probationary period.) Support program for self-learning • Up to 10,000 yen per month (before tax) is subsidized without prior application if you purchase books or courses that are necessary to achieve your mission. Congratulatory money to celebrate marriage and childbirth • When an employee gets married or his/her baby is newborn, we provide 30,000 yen to the employee as a congratulatory gift. Confidential Confidential 28
  29. 29 Communication across the company Various initiatives are taking place

    online to stimulate communication between members despite the remote work situation. ABECON • We hold internal study meetup on Fridays. • The presenters are our employees mainly. We sometimes invite presenters from outside such as specialists in certain fields, business owners of startup companies, venture capitalists, etc. • Participation is voluntary. There is no fixed topics and wee share knowledge on various topics. • We hold an external study meeting about once a month. (ABEJA Tech LT) We put information on connpass (https://abeja-innovation-meetup.connpass.com/event/228855/) Presentation by ABEJA members Presentation topics (examples) Data science, IoT, Deep learning, Art, Music, Sociology, Liberal arts, Law, Origami with mathematical approach, Art x Technology, Logical writing Presentation by external specialists Mathematics, SQL for business, Problem solving skills for startup business owners, Hardships in the course of success, Topology Asakai - An all-staff assembly every morning • Every morning, we meet up online. The designated member of the day presents for 3 minutes to share any topics such as their recent thoughts, lessons learnt, news, etc. Confidential Confidential 29
  30. 30 Guidelines for side-jobs Employees must apply for approval for

    the following side-jobs Activities conducted by ABEJA’s employees that fall under any of the following conditions: 1. Taking a position of director or officer of a corporation or other organization other than ABEJA, or entering into an outsourcing contract or employment contract with a corporation or other organization other than ABEJA.(*2) 2. Conducting business on their own 3. Conducting any other continuous commercial activities Confidential Confidential 30 At ABEJA, employees are allowed to have a side job because we believe being active and exhibiting their great talents in a different field from their regular workplace shall bring about positive impact on the employees themselves as well as the company and society. (*1) *1: Employees must obtain approval by the company in advance. If the company deems the side job is causing serious negative impact on performing the main job at ABEJA or the employee’s health, the company may consult with the employee to revise the side job or cancel the approval. *2: In principle, we do not permit employees who have employment contracts with ABEJA to enter into other employment contracts with other companies.
  31. Confidential 31 External communications Everyday life of technopreneurs (note) We

    introduce daily life of technopreneurs at ABEJA. https://note.com/abeja At ABEJA, we value technology and liberal arts. We share our stories and technologies in relation to our values through our blog articles. ABEJA Tech Blog We post articles on technologies we use in ABEJA. https://tech-blog.abeja.asia/ Confidential 31
  32. 32