Can AI ever completely replace translators? Not really. AI translations often lack the lexical richness their human-translated counterparts, mainly because AI models make decisions based on probabilities – not lived experience. Of course, AI can produce “accurate” translations, however the translations lack the spice of life, like a textbook version of the source text.
For many corporations, accuracy is all that’s required, which makes the AI translation sector attractive. SmartcatFounded in 2016, the corporate is one in all the providers of automated translation tools for enterprises and co-founder and CEO Ivan Smolnikov says business goes well.
“We have over 1,000 enterprise customers, including 20% of the Fortune 500,” he told TechCrunch. “While most of Smartcat's customers are large global corporations, we also count many local and international government entities amongst them.”
Before founding Smartcat, Smolnikov was a physicist on the Russian Academy of Sciences, where he researched fiberglass materials. After two years within the lab, Smolnikov decided to try his hand at entrepreneurship and founded the language services company ABBYY LS.
In fact, at ABBYY, Smolnikov founded Smartcat, which was spun off as a separate company in 2016. That same 12 months, Smolnikov left ABBYY, where he was then a member of the board of directors.
“I founded Smartcat to reinvent the normal translation agency model based on a time-consuming, operated by hand, long supply chain built on human services,” said Smolnikov. “Our AI platform supports a big selection of use cases for enterprise clients working with multilingual content.”
At a high level, Boston-based Smartcat offers tools, apps and managed services that help corporations translate written and spoken content – resembling e-learning courses, web sites, files and software – into around 280 languages. Smartcat doesn’t necessarily train AI translation models itself, but runs content through a “matching engine” that determines which third-party model may be suitable for the content and the goal output language.
Smartcat says it optimizes translation models – including internal models trained by customers themselves – where it is sensible, resembling in cases where an organization has a big database of steadily translated phrases that a model should “memorize.”
Smartcat recognizes that AI sometimes makes mistakes, so it also offers access to a network of translators and proofreaders who may help with proofreading for a fee. “Customers can make a choice from different translation options, resembling AI translation, human translation, or a mix of automatic translation and skilled proofreading,” Smolnikov explained.
For a couple of years now, some translators have been on Reddit advisable that Smartcat had an issue with dishonest – and non-paying – clients. However, Smolnikov assured me that this was resolved and that translators can charge whatever they need, minus Smartcat's 2% to eight% fee per payment. (The exact fee relies on the amount of orders processed and the countries of residence of the clients and translators.)
Although there are many corporations offering AI translation services, including those with human assistance (see: EasyTranslate, D-ID, DeepL, Lilt, Lengoo, etc.), Smolnikov believes that Smartcat primarily competes with traditional translation agencies and in-house organizations.
“Traditionally, corporations depend on outsourcing to agencies,” he said. “Insourcing was one other approach, but when done manually, it had issues with scalability… Smartcat's deal with the standard of voice AI offers a practical (alternative).”
Investors appear to agree.
Smartcat announced Tuesday that it has raised $43 million in a Series C funding round led by Left Lane Capital, bringing the corporate's total raised to $70 million. Smolnikov says the brand new money will go toward expanding the 200-person team, product development and various ongoing marketing and sales efforts.
Vinny Pujji, managing partner at Left Lane Capital, said in a prepared statement: “As a primary mover out there, Smartcat has a broad customer portfolio and is uniquely positioned to extend the depth and quality of its product offering to proceed to secure competitive advantage over time.”
These “advantages” could translate (no pun intended) right into a sizable payday should the rosiest predictions concerning the AI translation sector come to pass. After According to Grand View Research, the worldwide machine translation solutions market – valued at $978.2 million in 2022 – could grow at a compound annual growth rate of 13.5% from 2023 to 2030.
