Out-of-Vocabulary (OOV) Words are words or phrases that are not included in a speech recognition model's vocabulary or training data. These words are often difficult for AI systems to recognize accurately.
Voice AI platforms reduce OOV errors by expanding vocabularies, using contextual language models, custom lexicons, and domain-specific adaptation. This is especially important for brand names, technical terms, and industry jargon.