Toppan Printing develops AI review and proofreading services for finance

On October 18th, Letterpress Printing will use AI technology to learn industry- and company-specific notations and technical terms related to printed matter and digital media, and to proofread and proofread sentences according to the standards of each company. Announced that it has developed a “review / proofreading support system”. Since the same month, it has been provided and sold as a service to the financial industry such as banks, cards, and life and non-life insurance companies nationwide.

This system was developed based on the review and proofreading know-how that Toppan Printing has cultivated through the production of a wide variety of printed matter such as pamphlets and e-mail newsletters and digital media, and has been improved and commercialized based on the results of demonstration experiments. ..

  • System usage image

    System usage image

The three main features are “Detection of typographical errors using AI”, “Management of production regulations”, and “Optimal interface for production work”.

In typographical error detection using AI, AI learns notations and technical terms specific to the industry and introducing companies to detect typographical errors according to the standards required by each company, and also “Tenioha”. Misuse, misconversion of kanji, typographical errors, etc. can be judged and detected according to the context.

Regarding the management of production regulations, the system manages the rules and standards of media production, and by preventing quality variations between different producers and managers, it is possible to manage multiple different regulations for each medium.

Regarding the interface, by providing an interface that matches not only the check of the completed document by the confirmer but also the sentence check from the production stage, from the check at the time of production to online deficit entry and correction instruction issuance. It is said to support the digitization and efficiency of the entire flow.

The initial cost starts from 5 million yen and the operating cost starts from 500,000 yen per month. Going forward, the company will expand sales of this service to financial institutions, aiming to introduce 20 companies by 2020 and to achieve cumulative sales of 10 billion yen including related orders by 2022. In addition, we plan to develop functions and learn from AI for industries other than financial institutions, and expand to various industries and industries.