DeepL Voice preferred by 96% of professional linguists, outpacing leading competitors in spoken translation speed and accuracy
New study reveals DeepL Voice delivers higher-quality translations and more stable live captions in meetings across 14 language combinations, compared to Google Meet, Microsoft Teams and Zoom COLOGNE, Germany, March 25, 2026 /PRNewswire/ — DeepL, a global AI product and research company, today reveals results from an independent benchmark study evaluating the quality and stability of…
New study reveals DeepL Voice delivers higher-quality translations and more stable live captions in meetings across 14 language combinations, compared to Google Meet, Microsoft Teams and Zoom
COLOGNE, Germany, March 25, 2026 /PRNewswire/ — DeepL, a global AI product and research company, today reveals results from an independent benchmark study evaluating the quality and stability of real-time AI-translation and captions for meetings across widely used collaboration platforms; Google Meet, Microsoft Teams and Zoom.
DeepL logo
As real-time multilingual meetings become routine for global teams, enterprises increasingly rely on AI voice technology to support collaboration, customer negotiations and cross-border decision-making.
Commissioned by DeepL and carried out independently by Slator – the leading source of research and market intelligence for translation, localization, interpreting, and language AI – the study finds that DeepL Voice achieves the highest scores for both translation quality and caption stability, outperforming built-in caption translation tools available in Google Meet, Microsoft Teams and Zoom.
In high-stakes settings – from sales discussions and supplier negotiations to internal strategy meetings – even minor translation errors or unstable captions can significantly slow momentum, create confusion or undermine critical business decisions. As global business accelerates, AI-driven translation is no longer a convenience feature, it has become essential infrastructure. The strength of that infrastructure depends on two critical factors: the quality of AI translation and the stability of live captions.
Key findings include:
DeepL Voice leads translation quality in human evaluation, achieving a 96.4/100 quality score for DeepL Voice for Zoom and 96.3/100 for DeepL Voice for Teams, compared with 87–89 across other evaluated platforms.
DeepL Voice significantly reduces high-severity errors, lowering the rate of critical or major translation errors by 76% on average versus other platforms.
DeepL Voice produces fully passing translated segments 79% of the time, compared with 42% across competing tools.
DeepL Voice improves caption stability, earning stability scores of 88.6/100 (DeepL Voice for Zoom) and 85.8/100 (DeepL Voice for Teams) in Slator’s automated frame-level analysis.
DeepL Voice reduces caption churn (on-screen rewrites/flicker) by 37.6% on average versus Microsoft Teams and 54.7% on average versus Zoom.
Across all blind evaluations, DeepL Voice is preferred by 96% of professional linguists benchmarking the solutions.
“Language AI is becoming the core infrastructure for how global businesses operate. In that context, accuracy and stability aren’t features, they’re requirements,” said Jarek Kutylowski, CEO of DeepL. “This independent benchmark shows DeepL Voice delivering on both and at a level that sets a new standard for real-time communication. When professional linguists overwhelmingly prefer one solution, it’s a clear signal of where the market is heading.”
“We looked beyond basic accuracy scores and focused on the ‘human’ side of AI captioning: readability, fluency, and stability. We didn’t just want to know if the words were right at the end; we wanted to see how they behaved while a person was trying to read them. That meant evaluating both linguistic quality and how captions behave on screen and that’s where we can clearly see a new benchmark being set by DeepL,” said Alex Edwards, Head of Consulting, Slator.
Why stability matters as much as accuracy Slator’s report finds that even when translations are broadly accurate, frequent caption rewrites can interrupt comprehension and reduce usability in real-world meetings. To reflect how captions appear to end users, Slator measured stability using frame-by-frame analysis of rendered captions on screen, capturing flicker, oscillation and rewriting behavior over time.
Methodology Slator commissioned 28 professional linguists to conduct a blind evaluation of AI-translated captions across 14 language combinations (7 into English, 7 from English). The report benchmarks standard, out-of-the-box caption translation settings in Google Meet, Microsoft Teams and Zoom, and compares them with DeepL Voice for Microsoft Teams and DeepL Voice for Zoom, including product features available to end users (such as glossaries and recognition of spoken terms). Slator states that its conclusions are independent and that it retained full editorial control over methodology and findings.
Looking ahead to April These findings are released ahead of DeepL’s global spring launch on 16 April 2026, when the company will unveil major updates to DeepL Voice alongside expanded capabilities in translation automation and cross-platform collaboration.
About DeepL DeepL is a global AI company building the language infrastructure that powers global business. More than 200,000 business teams and millions of individuals use DeepL’s Language AI platform to communicate globally, collaborate and operate across languages in real time. By combining breakthrough AI models with enterprise-grade security and privacy, DeepL enables organizations to work seamlessly across markets and cultures. Founded in 2017 by CEO Jarek Kutylowski, DeepL now has more than 1,000 employees and is backed by leading investors including Benchmark, IVP and Index Ventures. Learn more at www.deepl.com.
About Slator Slator is the leading source of research and market intelligence for translation, localization, interpreting, and language AI. Slator’s Consulting practice is a trusted partner to enterprise buyers, language technology platforms, and language solutions integrators looking for independent industry analysis.
