July 2022 Summaries

This July, Speechmatics released its latest product update, featuring improved accuracy across major languages, an updated Advanced Punctuation feature to support readability, and the addition of a Finance Domain Language Pack for English, which provides accuracy improvements for financial terminology. The company has also improved the accuracy for French and Portuguese, added Ukranian as its 34th supported language, and updated 16 languages to support additional punctuation marks. Additionally, Speechmatics introduced its Language Identification Container, allowing users to identify the predominant language in any media file, and is now available through a Docker Container with support for 12 languages. The update also includes improvements to Speaker Diarization, including increased performance and a new sensitivity parameter. With this release, Speechmatics aims to serve the financial industry and expand its offerings into other industries.

Our latest uplift across French and Portuguese has resulted in significant improvements to our models, particularly noticeable in Canadian French and Brazilian Portuguese ASR, with average relative reductions in Word Error Rate of 18% for Canadian French and 5% for Brazilian Portuguese, showcasing a huge leap ahead of competitors like Google and Microsoft, as we continue to use self-supervised learning to improve on every language continuously.

Speechmatics has further enhanced its speech recognition technology by improving Canadian French and Brazilian Portuguese, two commonly used language variations with subtle and significant differences from European French and Portuguese. The company's industry-leading engine now caters specifically to these nuances, enabling unprecedented accuracy in speech recognition for end users operating in these languages. This improvement is a result of training the engine on 1,100 hours of Canadian French voice data, drawing from publicly available sources, which has led to an average relative reduction of 18% in Word Error Rate (WER) for its Standard model and 13% for its Enhanced model. The enhancements will allow users in industries such as contact centers to make the most of their speech recognition technology, where accurate communication and understanding are vital.

Speechmatics has finalized its Series B funding to the tune of $62 million, which will enable global expansion and improve infrastructure, as well as add firepower to research and development. The funding was led by Susquehanna Growth Equity, with participation from existing investors AlbionVC and IQ Capital, demonstrating faith in the organization's ability to deliver on promises. This investment will support Speechmatics' continued growth and innovation, particularly in its speech recognition engine, which has already achieved significant milestones, including a Queen's Award for Enterprise Innovation.

Speechmatics has launched an English language pack specific to the finance industry, designed to improve accuracy in speech recognition for compliance, fraud identification, analytics, financial news and earnings calls. The company's speech-to-text engine can now identify finance terminology, reducing errors by 40% and avoiding confusion with abbreviations, acronyms and finance-specific terms. This addition aims to support inclusivity in the financial services sector, where industry jargon can be confusing for standard speech-to-text engines. By leveraging its trained model on 200,000 hours of earnings calls transcripts, Speechmatics has overcome significant challenges traditional speech-to-text engines have struggled with. The company's new offering sets a blueprint for high-stakes industries with complex terminology, paving the way for industry-specific packs in sectors such as medicine and law.

Inclusivity is a key concept that benefits various industries by increasing audience reach, customer satisfaction, and compliance with international laws and standards. Localizing media content through Automatic Speech Recognition (ASR) supports languages spoken by over half the world's population, enabling diverse customer bases to access services without language barriers. Building inclusive classrooms and contact centers also promotes diversity and inclusion in education and customer service, respectively. By leveraging speech translation technologies, organizations can ensure compliance with international laws and standards while providing high-quality services to a broader audience.