Translated from TechBlitz. Original article can be found here.
The software that integrates speech recognition, machine translation, and subtitling.
Translated subtitles are indispensable in delivering video content globally, however, creating them entails numerous challenges.
As manifested by its motto, “Make your content accessible to the world,” AI Communis strives to facilitate video content localization and resolve such issues.
The inception of the company can be attributed to a rendezvous between Mr. Nobuhiko Suzuki, who previously worked as a venture capitalist in the AI sector for an investment firm, and Mr. Ian Lane.
“I was previously stationed in Singapore while working for a bank, after which I joined an investment firm. During my encounter with Ian, who was then a CTO at an AI start-up and also an associate professor at Carnegie Mellon University, we connected well due to our age proximity. Ian revealed that Southeast Asian international students who graduate do not often return to their home countries but rather secure jobs at Big Tech, resulting in the lack of specialists in natural language processing or speech recognition within their regions. Given my experience in banking and living in Singapore, I assumed I was able to contribute to business development. I had faith that combining my skills with Ian’s technology would result in a new business, hence why I began secondary research.”
His initial business concept involved a compliance-as-a-service system that records conversations at bank counters and analyzes them through AI. However, this idea wasn’t as successful as anticipated. Half a year later, he pivoted and launched a crowdfunding transcription service that aimed to raise funds and accumulate AI training data, which ultimately led to the development of Auris AI.
Although transcription is a low-cost venture and not especially lucrative, he discovered a demand for translated transcriptions and subtitling videos while establishing client relationships. It is typical to engage separate companies for transcription, translation, and subtitling, but he realized that by combining all these activities, they could establish a profitable enterprise. Therefore, he developed Auris, a software combining speech recognition, machine translation, and natural language processing.
By using Auris, the need to transfer data and outsource work to third-party firms has been eliminated, reducing the time required to create translated subtitles substantially. Mr. Suzuki introduced the software after recognizing its potential as a business, subsequently Mr. Suzuki and Mr. Lane founded AI Communis in Singapore.
Mr. Suzuki confirms that Auris was constructed with multilingual support in mind. We developed it to meet the need for translating videos of new product launch events held by American firms and adding Japanese and Chinese subtitles to them, as well as the need for adding translated subtitles to videos submitted to pitch competitions, thereby expediting the review process.
Proprietary Auris software and corporate translation solutions as revenue drivers
AI Communis generates revenue from two sources: paid Auris users and corporate solutions such as translation and subtitling services. Mr. Suzuki explains the correlation between the development of Auris and these solutions.
Auris operates on a freemium model. Users can register and access our transcription, translation, and subtitling services for up to 30 minutes per month free of charge. Beyond that, a paid package is required. As of early January 2023, they have a total of 122,000 paying and non-paying users. By utilizing our pool of freelancers, which consists of Auris users, AI Communis provides solutions for businesses.
Currently, approximately 3,000 Auris users are registered as AI Communis freelancers who are willing to accept work. By assigning these users with translation and subtitling projects commissioned by businesses, AI Communis can offer fast turnaround times, affordable prices, and high translation quality.
In addition to translation companies, AI Communis offers translation and subtitling solutions to entertainment companies and manufacturers seeking to expand their business overseas. About 70% of all translation tasks involve English to other languages. In terms of geography, India has the largest number of Auris users, followed by the Philippines, both of which have a sizable population of native English speakers.
“From the U.S. perspective, countries like India and the Philippines are on the ‘other side of the world,’ so the time difference is around twelve hours. This means that if you request a translation in the U.S. in the evening, it can be completed and delivered the following morning,” Mr. Suzuki notes.
Thorough knowledge of the translation and subtitling business
Strength in being able to meet the need for speed
Due to its superior performance and the ability to harness the talent of freelancers from its user base, Auris has a competitive edge in terms of reducing man-hours and delivery times. Mr. Suzuki has analyzed the origins of these strengths.
“To begin with, I have personal experience in translation and subtitling. While developing Auris’ features, I provided direct instructions to engineers from a translator’s perspective, combining AI with practical knowledge of the work. Additionally, my practical experience proves valuable when we receive corporate requests for translation and subtitling. Typically, corporate clients are not familiar with the subtitling workflow.”
“In addition to my practical experience, I possess expertise in speech recognition, machine translation, and natural language processing, which enables me to approach customer requests from a technical standpoint. Our strength lies in our ability to demonstrate how the combination of practical work and AI technology aligns with the needs of our clients.”
“Auris excels at meeting the translation requirements of recent years. In modern times, translation and subtitling needs have been divided into two categories: movie subtitles necessitate a high level of artistry and quality, while videos from YouTubers and other sources, which are frequently published, require fast turnaround times and competitive pricing.”
“By industry standards, subtitling a one-hour video used to take around 20 days, but we can accomplish this in roughly half the time, meeting the demand for quick delivery and affordable prices. This is yet another one of our strengths. Environmental factors that contribute to our success include the rapidly expanding video market and a dearth of competitive services in Southeast Asia.”
Seeking a partnership with an advertising agency responsible for the promotion of large companies
AI Communis aims to establish a business partnership or collaboration with a leading advertising agency responsible for producing promotional videos and developing marketing campaigns for large corporations. One such partnership could involve translating video content, which is crucial for the overseas expansion of major entertainment companies.
Mr. Suzuki highlights, “Apart from website translations, the demand for translating live videos and music videos has also emerged. In a potential partnership with an advertising agency that creates advertisements for large companies, we would assume the role of a business partner responsible for all translation requirements.”
Starting in December 2022, the Auris service will be available in Japan to further expand its user base. Mr. Suzuki, with a target of 300,000 users by the end of 2023, declares, “We exceeded 110,000 users at the end of last year and reached 122,000 in early January this year. We expect to achieve our goal by the end of 2023.”
Our primary focus is to increase the number of paying users. Mr. Suzuki outlines the necessary measures to achieve this.
“We aim to expand our payment options to include more local payment methods commonly used in Southeast Asia. Currently, we only accept credit cards, which can be a barrier for users in the region where obtaining bank accounts or credit cards can be more challenging compared to the US, Europe, or Japan. By incorporating payment services that are familiar to our local users, we can simplify the payment process and provide better service. While implementing multiple payment services can be time-consuming and challenging, we consider it an essential part of our value package to better serve our users.”
“In addition, we have plans to introduce a multilingual dubbing feature that utilizes AI-synthesized voice based on the translation text data generated through our subtitle business. In August 2022, we signed a contract for academic guidance with Shinnosuke Takamichi, an Assistant Professor at the University of Tokyo, who specializes in speech synthesis. With the integration of speech recognition, natural language processing, and speech synthesis, we aim to offer a one-stop solution for content creators. We will continue to address the growing demand for dubbing services in the future.”
“Similar to translation and subtitling, the voice-over market is bifurcated into two categories, one emphasizing quality and the other focusing on speed and affordability. Traditional voice-over services prioritize quality dubbing, but this process is time-consuming and costly, necessitating a dedicated recording studio and voice actors.”
“By utilizing a synthesized voice, we can offer a distinctive value proposition to our clients by ensuring the fulfillment of the speed and affordability requirements through advanced technologies. An instance of this is synthesizing the voice data from Japanese YouTubers with their consent, thereby allowing us to effortlessly dub their videos to make it appear as if they are speaking in a foreign language. Initially, we plan to offer B2C dubbing services to meet the demand for avatar and VTuber dubbing in the metaverse, with the ultimate aim of developing it into a B2B SaaS service in the future.”
About AI Communis
AI Communis, a Singapore-based startup, specializes in developing and managing Auris, an AI-powered software that automatically transcribes and translates various languages. Since its inception in November 2021, Auris has successfully garnered 120,000 users primarily from Southeast Asia and currently supports 18 languages, including but not limited to English, Japanese, Chinese, Korean, and multiple Southeast Asian languages. The company is now preparing to launch its operations in Japan by December 2022 with the primary objective of expanding its user base. We had the opportunity to speak with Mr. Nobuhiko Suzuki, the co-founder and CEO of AI Communis, about the company’s business model and future prospects.
About Nobuhiko Suzuki, Co-founder & CEO
With 18 years of experience in the finance industry as a corporate banker, senior management staff, and capitalist, Mr. Suzuki holds a Bachelor of Laws from the University of Tokyo and an MBA from Duke University. He relocated to Singapore in 2019 and founded AI Communis in April 2020.
Leveraging his practical experience as a freelancer on crowdsourcing websites in transcription, translation, and subtitling, he developed Auris AI, a software for translation and subtitling that employs natural language processing and speech recognition.