Would You Invest in a Translation AI Startup

2024 was a most interesting year for the language industry. AI translation has irreversibly changed the way services are provided, and as has happened any time a disruptive technology takes hold of a market segment, there have been shifts for LSPs and buyers alike, with some bankruptcies but also a healthy amount of M&As and investment rounds. 

More organizations, public and private, are incorporating AI into various business areas, including language access — adapting/adopting, or partnering with language technology companies. An increasing number of companies are also outright integrating AI translation as a feature (TaaF) into their existing products via large language models (LLMs). 

We asked readers about their business outlook for 2025, and close to a third (29.8%) of the respondents are optimistic and see it as positive. For almost a quarter (24.6%) things could be slightly negative this year, while the rest of the readers see their immediate future as very positive (15.8%), neutral (15.8%), or very negative (14.0%).

Would You Invest?

Slator research shows there is now a stronger M&A deal flow than in 2023, and motivated sellers are making it a buyer's market. Startups are receiving increased funding, particularly those with unique offerings in niche areas. While AI is disrupting the industry, it is also creating new opportunities. Companies successfully integrating AI and demonstrating feasible business models are attracting investment.

So would readers put their money where their work is? We asked readers as 2025 began if they would invest in an AI Translation startup. Opinions were spread out, with two identical cohorts (22.5% each) saying that it depends or probably not. An additional two equal-size groups (17.5% each) answered “of course” or “too many eggs in that basket already.” The rest (20%) would never make such an investment.

AI’s Resolute Pace 

To a degree, measured in terms of the risks implied and any quality assurance guards in place, government agencies could greatly benefit from AI written and speech translation. In the US, for example, multiple states are leveraging OpenAI and other options, including Pocketalk, in settings like education, unemployment, small business, and others.

In the UK, where for a few years now the Ministry of Justice has been scrutinizing the way translation and interpreting services are provided in the Courts and Tribunals System, the use of AI for court interpreting was part of the most recent debates on the ongoing inquiry.

Lord Porter of Spalding argued during a session on the subject in Parliament that if the health service was starting to embrace the technology, the courts should too, stating that the consequences of using AI in that setting are “not as dire as they would be in the health service.”

We asked readers how they think the use of AI in court interpreting should indeed proceed, and the majority (66.7%) chose the option to use caution. Less than a quarter (21.6%) opted for the more cautious approach to launch a pilot first, and the rest of respondents (11.7%) would actually start using the technology immediately.

New Lingo Taking Hold

AI has regaled us with quite a few new terms, and although it is not easy to rank them in terms of importance across all language applications and individuals, some terms are decidedly more frequently used and impactful than others, such as “agentic machine translation.”

As 2024 came to a close, we looked at the terms that were more influential in the narrative around language AI throughout the year, and decided that “agentic machine translation” was indeed on the top five, along with “translation as a feature” (TaaF), Retrieval Augmented Generation (RAG), “prosody,” and the still hotly debated “Expert-in-the-Loop,” whose role continues to morph with the technology. 

We asked readers to rank these terms, and Expert-in-the-Loop garnered the most votes (34.5%). Translation as a Feature (TaaF) was a close second (30.9%), while agentic machine translation was the top term for about one in six poll respondents (16.4%). Retrieval Augmented Generation (RAG) was the top term for a tenth (10.9%) of readers and the rest chose “other” terms (7.3%).