Tech company Intento raised USD 3m in seed round funding led by Flint Capital, with the participation of Berkeley SkyDeck Fund, Smarthub, and angel investors. The news was first shared in a press release on Slator and Crunchbase News. Combined with the initial USD 1.3m the company has raised since its 2016 founding, Intento has raised a total of USD 4.3m to date.
The Berkeley, California-headquartered company trains, evaluates, and manages custom and stock neural machine translation (NMT) engines for, so far, about 30 clients in the global retail, travel, and technology industries.
Intento’s AI integration platform, Intento Enterprise MT Hub, offers a single API that connects to multiple MT models, directing clients’ requests to the most appropriate MT engine on a case-by-case basis. (Intento was recently granted a patent for the technology behind the API.)
Co-founders Konstantin Savenkov and Grigory Sapunov, who hold PhDs in Computer Science and AI, respectively, started Intento to modernize the way businesses carry out AI initiatives.
Savenkov, Intento’s CEO, left academia in 2008 to write content recommender systems for music streaming and books. Prior to founding Intento, he was COO of the subscription-based ebook app Bookmate. CTO Sapiunov’s past experience includes working as a tech lead at Yandex News and developing core AI technologies for multiple startups.
SlatorCon Remote June 2022 | Super Early Bird $98
A rich online conference which brings together our research and network of industry leaders.
Although Intento works with other AI services, such as image tagging, transcription, and sentiment analysis, Savenkov told Slator, “We decided to start with machine translation as this is a universal need for any large company.”
“The emerging MT curation industry will overtake the language industry at some point” — Konstantin Savenkov, CEO, Intento
Intento soon found that an API connecting to multiple MT systems was not enough, as clients requested guidance selecting the best MT engines for their needs. The company responded by developing tools for data cleaning and model evaluation.
By 2019, Savenkov said, several of Intento’s large clients had begun moving multi-vendor MT into production. Using five or six MT engines across up to 10 enterprise systems highlighted the limits of the clients’ existing tools, prompting Intento to eventually release its AI integration platform.
“Most of the scenarios are real-time translation, such as live chats, support tickets, knowledge bases, and on-the-fly website translation,” Savenkov said.
Valuation was not disclosed, but Intento recently went over USD 1m ARR, according to Savenkov. The CEO told Slator that Flint was selected as a lead investor in this funding round due to Flint’s experience in Enterprise SaaS and venture networks.
Savenkov said that Intento plans to use these funds on Sales, Marketing, and Customer Success, working with managers of localization, customer support, community management, and content management.
According to Savenkov, Intento has seen a spike in interest from large enterprises in light of Covid-19. The pandemic has prompted more work to move online and has required rapid cost-cutting, both of which have fostered an open-mindedness to new methods and digital transformation.
Looking ahead, Savenkov predicts that the current trend of unbundling data, technology, and computing ownership will continue to shape the MT landscape. He added that, coupled with domain adaptation, this unbundling enables data owners to both train and monetize their own models.
On the technical side, full-document and context-aware translation, as well as improved tools to control translation style, Savenkov said, will bolster the “emerging MT curation industry, which I believe will overtake the language industry at some point.”