Amazon Bedrock Data Automation now supports custom vocabulary

Amazon Bedrock Data Automation (BDA) now supports custom vocabulary through a new capability called Data Automation Library, enabling you to improve speech recognition and transcription accuracy for domain-specific terminology when processing audio and video content. BDA automates the generation of insights from unstructured multimodal content such as documents, images, audio, and videos for your GenAI-powered applications. This new capability is designed for organizations in healthcare, legal, financial services, media, and contact center analytics that need accurate recognition of specialized terminology.

With custom vocabulary, you can provide BDA with lists of domain-specific words and phrases—including brand names, acronyms, medical terms, and technical jargon—so that these terms are accurately recognized and extracted from your content. You can also specify display forms to control how recognized terms appear in output, such as displaying "electrocardiogram" as "ECG" or "discounted cash flow" as "DCF". For example, a healthcare company processing doctor-patient conversations can provide lists of disease and drug names, or a contact center analytics provider can supply industry-specific terminology across verticals such as insurance, telecom, and public sector to ensure accurate transcriptions. Custom vocabulary supports 11 languages including English, Spanish, French, German, Italian, Portuguese, Japanese, Korean, and Chinese (Simplified, Traditional, and Cantonese), and is included at no additional charge.

Data Automation Library with custom vocabulary is available in US East (N. Virginia), US West (Oregon), Europe (Ireland), Europe (London), Europe (Frankfurt), Asia Pacific (Mumbai), and Asia Pacific (Sydney). To learn more, visit the Bedrock Data Automation page or view the Bedrock Data Automation User Guide.