Solutions Engineer

Speechmatics has launched a bilingual Arabic–English speech recognition model that handles both languages simultaneously, including the world's first bilingual medical model for clinical settings. The company claims 35% lower word error rate than Google on code-switching tasks and 24% better performance on Arabic-only transcription. The model supports Gulf, Egyptian and Levantine Arabic dialects and can be deployed on-premises, on-device or via cloud, addressing data sovereignty requirements across MENA. It includes speaker diarization, real-time streaming and batch transcription capabilities. The medical variant is trained on twice the vocabulary of Speechmatics' English medical model, accurately transcribing drug names, procedures and clinical terminology across both languages. Healthcare AI company Sully.ai has adopted the technology for its MENA expansion.

Speechmatics has partnered with Edvak EHR to embed enterprise-grade speech recognition into clinical workflows. The collaboration enables Edvak's AI-native electronic health record platform to convert live clinical conversations into structured documentation, triggering automated tasks and care coordination directly within the EHR system. Speechmatics' English Medical Model delivers 93% real-time accuracy and 96% medical keyword recall, distinguishing between similar terms like hypertension and hypotension whilst preserving clinical meaning essential for coding. The technology addresses critical accuracy needs where a single dropped word can reverse documented meaning. The partnership supports Edvak EHR's shift from standalone transcription to embedded intelligence, with physicians able to review and adjust AI-generated outputs in real time. Speechmatics supports on-premises, private cloud and SaaS deployment meeting HIPAA-aligned compliance requirements.

Speechmatics and Boost.ai partner to power enterprise voice AI for Europe's most regulated industries. Two European AI leaders combine forces to deliver responsible, enterprise-grade technology for financial services, healthcare, and public sector. CAMBRIDGE, United Kingdom, Feb. 12, 2026 (GLOBE NEWSWIRE) - Speechmatics and Boost.ai today announced a strategic partnership to accelerate the deployment of enterprise-grade voice AI across Europe's most highly regulated industries. The collaboration brings together Speechmatics' speech recognition technology with Boost.ai's conversational AI platform, already trusted by leading banks, insurers, healthcare providers, and public sector institutions across the continent. Voice AI is no longer experimental. In financial services, healthcare, and government, it has become critical infrastructure. But most voice systems still struggle to reach enterprise standards of accuracy and control when deployed at scale. When speech fails to understand accents, dialects, or domain language, the impact is immediate: compliance risk, broken workflows, and customers excluded from essential services. This partnership focuses on production environments where accuracy, reliability, and control are non-negotiable. Where regulated AI is already working at scale Boost.ai has established itself as one of Europe's strongest forces in regulated conversational AI. Recently named a Leader in the 2025 Gartner Magic Quadrant for Conversational AI Platforms, the company has become the platform of choice for organizations that need automation to perform under sovereign regulatory pressure. Nine out of ten Norwegian banks now rely on Boost.ai, and 118 municipalities share a common voice AI platform. That success is now expanding across European banking, insurance, healthcare, and public services. Speechmatics was selected after extensive evaluation of how speech technology behaves in real environments. The decision was driven by language accuracy, real-time performance, and the ability to handle dialects, accents, and complex linguistic conditions that define European voice interactions. "We needed speech recognition that performs when the stakes are real. Speechmatics brings the accuracy, stability, and language coverage required to operate in regulated industries, including complex markets like the Nordics," said Samantha Rosendorff, VP of Go-to-Market at Boost.ai. Built for Europe's language complexity Speechmatics has invested heavily in Nordic and European language modeling, including Swedish, Norwegian, Danish, and Finnish. These languages present challenges that most speech systems struggle with: compound words, dialect diversity, and multilingual code-switching. Custom dictionaries, adaptive retraining, and domain-specific modeling ensure terminology is captured accurately every time. In regulated industries, if speech systems fail across accents or dialects, they fail operationally. Accuracy becomes a compliance requirement. Designed for production, not benchmarks The partnership delivers speech technology optimized for live, high-volume environments. Speaker Focus distinguishes between speakers in noisy contact centers to support accurate audit trails. ForceEndOfUtterance gives precise control over conversational timing, delivering final transcripts within 250ms. Advanced speaker diarization ensures correct attribution in multi-speaker scenarios. Flexible deployment across cloud, on-premises, and on-device environments supports GDPR and data sovereignty requirements. "Boost.ai has built one of the most trusted platforms for regulated conversational AI in Europe. Our role is to make sure the speech layer is strong enough to support that trust," said Katy Wigdahl, CEO at Speechmatics. "This partnership is about giving Boost.ai and its customers a speech foundation that holds up in their most critical systems." The partnership will now extend to the US market, bringing the same production-grade approach to American financial services, healthcare, and public sector organizations.