Tokyo Office Tour: Xenodata Lab. uses AI to help investors analyze financial...

Tokyo Office Tour: Xenodata Lab. uses AI to help investors analyze financial reports

SHARE

See the original story in Japanese.

This is part of our ‘Tokyo Office Tour’ series (RSS), a modest attempt to better understand how folks in the local startup scene are working every day.

We have started to see startups here and there renting houses around Tokyo to use as offices. Instead of rural or suburban areas, this is happening inside of the Yamanote loop train line where rental prices are comparatively high. When asked what they did before moving to their new office, they answer that they used members’ living rooms, or co-working spaces, or even the Renoir coffee shop (yes, the same Renoir where Japanese leading game developer Gumi got its start).

It is probably even the case that some of the empty rooms in the houses are being used as living spaces for team members. As long as they apply to divide the areas at the tax office, they should be able to cut costs for the area used as offices, and through the separation of public and private space we could see the birth of a new environment that combines living and working areas. The challenge of how to approach work-life balance would come up, but such a concept is foreign to a single entrepreneur looking to establish their company.

FinTech startup Xenodata Lab. recently rented a house between Ebisu and Hiroo in Tokyo to use as their base of operations. Prior to this, they had worked out of one of their member’s living rooms located in Roppongi, but are now ready to forge the foundations of their business from within their “new castle”.

Ebisu is a popular area for offices, and nearly ten years ago when I was doing business, had I rented offices there I remember the month rent per tsubo (about 3.3 square meters) did not drop below 60,000 yen (around $539 US). Often a 10 month security deposit is necessary for office spaces, making it difficult to use for a business, and of course, you must pay a no interest deposit of a few million yen (tens of thousands of US dollars) to the landlord upon renting. Using the nest egg collected from investors to pay for a deposit is something, as a proprietor, I’d like to see avoided, but in using a residential house as an office the security deposit becomes relatively cheap, significantly lowering the hurdle for finding a space to do business in.

Xeno Flash, developed by Xenodata Lab., specializes in listed stocks in Japan and uses XBRL (eXtensible Business Reporting Language) analysis, PDF table analysis, and PDF graph analysis to convert information attached to various financial-related materials into tabular data, which they then apply their own algorithm to in order to extract the most important financial points, and furthermore, through natural language processing, they are able to pull up a background of specific numerical values from an enormous amount of sentence data in the material.

Of the 3,600 listed Japanese stocks, only 500 companies, corresponding to 14% of all stocks, are issuing financial reports which individual investors can refer to when making investment decisions. In other words, since the price fluctuations are so intense, account analysis reports are not issued by most of the small and medium-sized companies invested in by individuals, but if they use Xeno Flash, it would be possible for companies to make them available.

Xenodata Lab. won the Grand Prix of MUFG FinTech Accelorator’s first batch last year. In February of this year, they raised 60 million yen (about $539K US) from the Bank of Tokyo-Mitsubishi UFJ, Teikoku Databank, Kabu.com Securities and Mitsubishi UFJ Capital in a seed round.

Currently, Xenodata Lab. is focusing on sales to online and offline retail brokerage fims as well as financial analysts in investment banking departments of securities, and in the coming months we may expect big announcements from them, such as large collaboration projects with securities firms. Founder and CEO Yojiro Seki confided he would like to expand the range of automated analysis for financial data to include unlisted stocks and foreign listings, and also increase their clients by 50 to 100.

Although the engineering personnel necessary for the immediate system development seems to be satisfied, since the company’s business requires expert knowledge of analysis algorithms and natural language processing, etc., they are always looking for talented people in this field. If you are interested in a business that might be able to create a big impact on the securities industry, try getting in touch with them.

Translated by Amanda Imasaka

XenoHouse situated in a quiet neighborhood
At first glance the signboard appears to be that of a cafe or hair salon.
CEO Seki relaxing on the rooftop
xenoHouse rooftop panoramic image
View from the rooftop overlooking Ebisu (click to enlarge)
The engineer team busy developing.
The winning plaque of the MUFG Accelerator is transparent making it difficult to capture in a photograph.