Sales Calls → Embedding Analysis → Lead Scoring → CRM Update

Automatically transcribe sales calls and extract key conversation segments, including customer pain points, budget discussions, decision-making timeline, and competitor mentions.

Process call transcripts through Embedist to create embeddings that capture the semantic meaning of customer responses. Compare these against high-value customer conversation patterns to identify buying signals.

Use GPT-4 to analyze the embedding similarities and extract specific scoring criteria: budget qualification (1-10), urgency level (1-10), decision authority (1-10), and fit score (1-10) based on the conversation content.

Use Zapier's built-in math functions to calculate a weighted composite score from the individual factors. Set up logic to categorize leads as Hot (80-100), Warm (60-79), or Cold (below 60) based on the total score.

Automatically update the lead record in HubSpot with the calculated score, individual factor ratings, and suggested next actions. Create tasks for sales reps based on the lead temperature and identified opportunities.