Get the Most Out of AI Linguistic Quality Assessment With ContentQuo

A

rtificial intelligence (AI) can speed up human and machine translation (MT) linguistic quality assessment (LQA) and allow organizations to assess more of their content with the same budget. However, in practice, integrating AI LQA into existing workflows and translation management systems (TMSs) can be difficult. Additionally, measuring the reliability of AI LQA output poses challenges even to large localization teams.

That’s why ContentQuo, which develops automated LQA software, launched a platform meant to simplify and enable AI LQA deployment. The company’s co-founder and chief executive officer (CEO) Kirill Soloviev presented the new offering at LocWorld 52’s Process Innovation Challenge (PIC) in October 2024 and was awarded third place in the competition.

Soloviev explained how it works in his PIC presentation: “First, you set up virtual LQA assistants using any large language model (LLM) you have an application programming interface (API) key for… Or, you can plug in third-party AI LQA engines built by external vendors. Then, you can actually test those virtual assistants at scale against previously done human LQA.”

Through relevant metrics and reports, the platform allows localization teams to monitor AI LQA performance over time and thus determine which automated engine is best for a given language pair, product, or content type. Plus, the flexible solution is compatible with any TMS.

But the best part might be that ContentQuo’s platform is useful beyond just translation workflows — it can also assess LLM output used for monolingual business functions such as marketing and customer support. Soloviev believes that localization teams that own LLM quality across their entire organization can gain more influence internally.

“If we can step up our game and have the courage to get outside our bubble, we can take over — if not the universe — at least a few more seats at the table in our respective corporations,” Soloviev said.