Invoca’s new transcription engine is powered by our custom generative AI LLMs (large language models) and turns call recordings into text with unmatched accuracy. To be clear, this is NOT just off-the-shelf transcription technology! Our data science and engineering teams took best-in-class generative AI LLMs and spent the last year training them on 700,000 hours of supervised audio data. The result is transcripts with an industry-low WER (word error rate) a.k.a. “They’re really accurate!”
More accurate transcripts means:
- They’re faster and easier to read
- It’s easier to analyze and act upon the insights they deliver
- The increased accuracy has a halo effect on other features that use our transcription like Signal AI Studio, keyword spotting and smart alerts
Best-in-class transcripts will be available to all customers who have Invoca transcripts enabled.
