ORIX Corporation Achieve efficient form management by utilizing AI data. The secret to the success of ‘PATPOST’

Please tell us a little about your company
Orix Group is a group of companies that develops businesses in areas as diverse as corporate finance, industrial/ICT equipment, environmental energy, automobiles, real estate, business investment, concession, banking and life insurance. I belong to the Orix Digital Strategy Promotion Office which provides new services that use digital technology. I’ve been particularly involved in the PATPOST service, released in May 2023, which had 1,500 companies as users (as of July 2024, including free trials).
Thank you. Could you tell us a little about your career up to this point?
I joined Orix in 2018 straight from university and am now in my seventh year with the company. Starting off in sales, I then moved to my current department. After transferring in October 2023, I worked on the creation of invoice systems. With electronic transaction data storage scheduled to become fully obligatory under the Electronic Bookkeeping Act on January 1, 2024, I started working as the person in charge of PATPOST in order to create a problem-solving solution aimed at streamlining the document management work that can be a burden on small and medium-sized business accountants.

What kind of service is PATPOST?
It’s mainly aimed at accountants and provides online document management storage that streamlines compliance with the electronic transaction data storage requirements of the Electronic Bookkeeping Act* and the invoice system. (*In addition, it is also necessary to establish and comply with administrative processing rules regarding the prevention of correction and deletion.) All you have to do is upload the report and the AI-OCR will automatically extract the relevant information for you. While similar services from other companies are charged according to usage, we charge a standard rate of 980 yen per month per user (excl. tax, minimum contract period one year, minimum of three users) and this has been well-received by many customers.
So that’s where AI comes in! How did you hear about our annotation tool ‘harBest’?
When PATPOST was being developed, I asked a number of annotation companies for learning data and data with location information attached. However, the data we received never seemed to lead to an increase in accuracy. Then, someone from our company met someone from APTO at an event and introduced them to the members of our department who work on digital solution development.
So there’s proactive communication between departments?
We have a very open work culture. I heard that there was a tool which could provide high-quality data creation and collection quickly so I got in touch with APTO, explained what Orix was looking for and we agreed to work together. I loved how proactive everyone at APTO was and I felt that the service was very fairly priced. Because of APTO’s help, we were able to process data much more efficiently.
What were some of the details of the project?
PATPOST streamlines the processing of document data such as invoices, but in order to extract a wider variety of items, more invoice templates were needed. By using harBest, our collection and annotation has become more efficient, even within more difficult categories such as invoices and receipts from overseas.

harBest utilizes crowd workers for data collection and annotation tasks. Did this help in terms of speed and volume?
I’m very satisfied with my communication with APTO, including during our discussions about delivery and the advice I was given on how to collect data to make better use of it in the future. Going forward, I would like to further improve the accuracy of the solutions by making the item definitions more precise.

Thank you! Could you tell us about your plans for the future?
We want to keep PATPOST’s simple design while adding the latest developments in AI. We also have a growing need for support for overseas forms, and we aim to further expand our services by collaborating with APTO, especially regarding the handling of form data from Southeast Asia.
Thanks very much for your time! We will continue to do our best to improve and hope to provide an even better service in the future. We look forward to working with you!
関連事例
-
LLM development at the highest level in Japan. What are the challenges faced by a research team devoted to improving accuracy?
Institute of Physical and Chemical Research(RIKEN)
- IT/Internet
- R&D
- Annotation
- Data collection
- Data Management
- Experienced
-
Search real estate all over the world at once using satellite data. What kind of future will “WHERE” make possible?
Penetrator
- IT/Internet
- real estate
- Annotation
- Data collection
- Data Management
- Experienced
-
“I started developing AI behind the scenes at a television station. Now I want to spread this throughout the company”
FUJIMIC, Inc.
- Annotation
- Data Management/Labeling
- IT/Internet
- Annotation
- Data collection
- Experienced
-
Streamlining Development After Successful Outsourcing of High-Volume Annotation Work
The Ricoh Company, Ltd
- AI Development (Experienced)
- Annotation
- Data Management/Labeling
- IT/Internet
- Annotation
- Experienced
-
How Leading AI Vendors Handle Essential Training Data for Generative AI
LightBlue
- AI Development (Experienced)
- Annotation
- Data Management/Labeling
- IT/Internet
- Annotation
- Data collection
- Data Management
- Experienced
-
Micro Control Systems: AI “Visualizing” Factories to Enhance Manufacturing
Micro Control Systems
- AI Development (Experienced)
- Annotation
- Data Management/Labeling
- IT/Internet
- Annotation
- Data collection
- Experienced
-
harBest Boosts MiiTel’s AI Speech Recognition
RevComm Inc.
- AI Development (Experienced)
- Annotation
- Data Management/Labeling
- IT/Internet
- Annotation
- Data collection
- Experienced
-
Challenges of Developing “LHTM-2”, a Large-Scale Language Processing Model From Japan.
alt, Inc.
- AI Development (Experienced)
- Annotation
- Data Management/Labeling
- IT/Internet
- Annotation
- Data collection
- Experienced
-
AGRIST Interview: Developing Agriculture Technology to Support the Aging Farmers of Japan
AGRIST
- AI Development (Experienced)
- Annotation
- Data Management/Labeling
- Agriculture
- Annotation
- Data collection
- Data Management
- Experienced