Executives Say more than Words on Earnings Calls

---

Attributes such as sentiment, complexity, and transparency are scored.  Markers for deception and bias are quantified.

 Audio is analyzed at the sentence level to extract vocal delivery features such as pitch, intensity, and speech rate. These metrics surface changes in emphasis, confidence, and tension that transcript-only approaches do not capture. 

 Language and vocal delivery are analyzed together to surface divergences between content and delivery, often revealing changes in confidence or credibility not visible in transcripts. Results highlight specific passages and topics for further analysis.