Robotics process automation startup Cinnamon secures $8.3M in series B round

Cinnamon’s AI Lab members in Vietnam with CTO Hajime Hotta in the center of the front row.
Image credit: Cinnamon

See the original story in Japanese.

Japanese startup Cinnamon, developing a document reading engine for “white collar” businesses, announced today that it has secured 800 million yen (about $7.4 million) in a series B round funding and 100 million yen (about $926,000) in loans. The latest round is led by SBI Investment with participation from FFG Venture Business Partners, Itochu Techno-Solutions, Sony Innovation Fund and TIS. Loans were secured from Mizuho Bank and Sumitomo Mitsui Bank this time.

The latest round follows the previous funding from multiple institutional and individual investors back in January this year (the announcement was made in February). The sum of funding in the previous round has not been disclosed but Cinnamon revealed that it was part of a series A round in an interview with The Bridge. When Cinnamon changed its business model from developing photo-sharing apps to an AI (artificial intelligence) solution provider, their incorporated entity was also changed over from a Singaporean company into a Japanese one. Combined with the sum raised by the previous entity, this round saw Cinnamon having raised a total of over 1 billion yen (about $9.2 million) to date.

The company will use the funds to strengthen hiring and fostering AI engineers, expanding sales channels, solidifying team structure and develop new products. The company is managing, developing business and sell products based on five offices in Asia – Tokyo, Hanoi (Vietnam), Ho Chi Minh City (Vietnam), Taipei (Taiwan) and Singapore. Among all of them, Hanoi, Ho Chi Minh City and Taipei locations have “AI Lab” functions to foster engineers and develop technologies with an aim to employ 500 AI engineers by 2022.

A briefing and lecture given to AI engineers and potential employees
Image credit: Cinnamon

The company’s main focus is the RPA (robotics process automation) class 2 solution Flax Scanner, which uses the document reading engine “Cinnamon AI” also developed by the company. Instead of simple OCR (optical character recognition), the Convolutional Neural Network (CNN), which is one type of deep learning, finally achieves accurate document reading with a precision rate of 99% or more. The biggest feature is not only the document reading, but also a system that can understand the context of the document and capture information in a semantic manner that databases and other systems can easily handle.

Specific use cases include contracts, resumes, sales progress reports, medical charts, handwritten application forms, real estate property information, receipts, etc. In particular, it is expected to be utilized in the fields of finance and insurance where paper documents are abundant, and it set the general goals of quadrupling business speed and cutting costs to a quarter.

Among the investors participating this round, FFG Venture Partners (the investment arm of Fukuoka Bank) looks at synergy in streamlining document processing at banking operations while system integrators like Itochu Techno-Solutions and TIS are expected to provide the company with partner sales channels leveraging their vast network with potential corporate users.

Cinnamon is also developing Rossa Voice, a voice dictation solution based on an automated word correction technology used for Flax Scanner. Rossa Voice allows users to transcript records at call centers and correctly dictate proper nouns or technical terms in discussions between representatives and customers.

Edited by “Tex” Pomeroy