"In the era of big AI, machines will help humans simplify complex tasks and handle tasks that are beyond human reach." With 20 years of experience in text mining, Song Hao, general manager of Blue Planet Information Company, firmly believes , under the trend of "open data" (Open Data), open data is definitely a treasure worth exploring. Through text mining and AI technology, Blue Planet uses "ΣCOUT" to build a huge corporate network, helping banks and bidding agencies fully understand customers and supply chain manufacturers, and effectively reducing business risks.

According to statistics from the "2020 Small and Medium Enterprises White Paper", the number of small and medium-sized entrepreneurs in Taiwan in 2019 was 1.491,420, accounting for 97.65% of all enterprises, an increase of 1.72% from 2018, setting a record in recent years. In the face of numerous small and medium-sized enterprises and new start-ups, whether it is a bank that needs to "know your customer" (KYC) or a government bidding unit that needs to select suppliers, it is a very time-consuming and labor-intensive task. "

It takes an average of 12 hours for a bank to produce a due diligence report. The "actual report" can be shortened to 1 hour

After conducting on-site visits to multiple banks, Blue Planet Information found that in the past, when a bank conducted due diligence on a single customer, it needed to first collect dozens of information from the Judicial Yuan, the Ministry of Economic Affairs’ Industrial and Commercial Registration, the International Trade Bureau, and media news. The relevant information on the website is then sorted and reviewed until an audit report is produced, which takes a total of about 12 hours of labor costs. However, through the "Report ΣCOUT" business history inquiry system developed by Blue Planet, using automated technology, all verification matters can be completed with one click without any gaps, and the time to complete a report has been reduced from 12 hours to 1 hour. Save more than 80% of time costs.

▲The Blue Planet team is committed to saving customers the trivial time that technology can do for them.

As an information company officially transferred from National Taiwan University, Blue Planet was formerly the R&D team of National Taiwan University Information Technology Institute. Since 1996, in order to preserve long-standing and precious archives, the national government has cooperated with National Taiwan University to implement the "National Digital Collection Project" ”, using text analysis technology to digitize various collections stored in libraries, museums, art galleries, etc. The research quenching during this period has also laid a solid technical foundation for Blue Planet. "Fact Report ΣCOUT" contains more than 3 million public offering information, ranging from food stalls, studios, small and medium-sized enterprises, large enterprises, and even companies that have ceased operations and completed liquidation. Including the 15 million judgments of the Judicial Yuan, as well as various news, forums, communities and other information, a total of more than 100 million pieces of data are stored in the database.

Song Hao pointed out that because Blue Planet started out with Chinese text exploration, in addition to structured data, the team is better at unstructured data, such as news reports, referees, etc. Accurate analysis requires a considerable technical threshold, so they Having a competitive advantage, it is not easy for latecomers to catch up. Especially once the court's judgment documents involve enterprises and trade transactions, there are often huge business connections hidden in them. Deep learning and AI algorithms need to be used to turn them into structured data, in order to further uncover the hidden relationship links.

Business risks are everywhere "True Report" Automated AI algorithms significantly reduce risks

The customer base of "ΣCOUT" is mainly divided into three categories. The first category is the financial industry that needs to conduct due diligence and credit investigation on customers; the second category is government units that often invite external tenders; and the third category is the general private sector. The purchasing unit of the enterprise. Song Hao said, "Business cooperation risks are everywhere, and the risks are especially high when encountering small and medium-sized enterprises." No matter at home or abroad, there are a large number of small and medium-sized enterprises, and information is difficult to collect. In addition, international regulations on money laundering prevention and combating financial terrorism ( AML/CFT) requirements have certain standards. Whether it is financing or lending, banks need to conduct detailed checks on their customers through systems like "Report".

▲Use exclusive semantic analysis technology to extract key words from the judgment and reveal corporate risks.

In addition, the government procurement department can handle hundreds of bidding cases in a day; general enterprises have countless preparatory tasks when selecting procurement vendors. If they rely on manual inspection, it can be said to be time-consuming and time-consuming. If an automated system can be used to conduct a detailed investigation of the "net worth" of bidding and purchasing manufacturers, it will be able to reduce risks such as failure to perform contracts and work safety accidents in the future, and it will also be able to strengthen the reminder of whether they are good manufacturers or refuse to deal with them. Using AI intelligent algorithms, Blue Planet exclusively developed "Business Network Diagram" and "Ownership Structure Table" to help uncover the intricate network behind the enterprise. Song Hao further pointed out that taking Far East Group as an example, in the past, when conducting cross-shareholding analysis, bank due diligence personnel had to draw 200-300 relationship routes with bare hands. However, with "real report", it only takes a few seconds and is complicated. The cross-shareholding network can be seen at a glance.

▲Through the formalized process of data inventory, collection, and cleaning, key information is extracted to build a business network.

Another example is that there are many investment targets in the securities market. If investors do not understand the company, the investment risk is quite high. Through the "real report" business network diagram, they can smell out business clues, such as whether the company is What's more, the resurrected "shell companies" can also unearth the core figures and ultimate beneficiaries (UBO) hidden behind the related companies. Regarding the situation where the person in charge, directors and supervisors sometimes have the same name, Song Hao said that AI will give different weights based on the similarity of the company's activity period and activity industry. If the weight ratio is higher, the possibility of judging that they are the same person will also increase. high.

Taiwan Open Data is at the forefront of the world and Blue Planet expects to open up new fields and markets within 5 years

After many years of hard work, Blue Planet Information is now a leader in the field of text exploration in Taiwan. Song Hao pointed out that when former Executive Yuan President Zhang Shanzheng was a political councilor, he vigorously promoted "open government data" and Taiwan's Open Data was at the forefront of the world, comparable to the United Kingdom and Japan. Most of the Open Data information in China is text, while in European and American countries it is PDF files or scanned files, which cannot be easily textualized and even more difficult to add value to. Therefore, Blue Planet Information seizes the opportunity to collect data from the government’s open data platform and apply it in value-added applications, allowing AI to combine big data from different fields, and is expected to open up new fields and markets in the next five years.

In terms of expanding overseas markets, Song Hao revealed that a foreign bank had approached him in the past, saying that it had multiple bases around the world and hoped to collect corporate information from 30 countries around the world. Such an opportunity also allowed him to start I have been thinking about it, hoping to copy Blue Planet’s operation and service model in Taiwan to the world. Therefore, the next stage of the company's operations is to study the data that can be publicly collected around the world, further promote the service model to overseas markets, and create an international business performance database. . Since its establishment in 2013, Blue Planet Information has continuously deepened its technology and moved the laboratory's research and development results towards commercialization. Song Hao said that people who come out of the research laboratory usually have their own ideas and persistence in research, but after commercialization, they inevitably have to compromise with the reality. He hopes that in 10 years, Blue Planet can be pushed to the road of public offering (IPO). , leading the team to become a technology leader and achieving the ultimate goal of becoming the "Light of Taiwan" in the software industry.

▲Dr. Song Hao, General Manager of Blue Planet Information

「Translated content is generated by ChatGPT and is for reference only. Translation date：2024-05-19」