What problem does it solve? Running sales conversations on WhatsApp manually is slow and inconsistent: leads go unanswered, objections get improvised responses, and CRM records fall out of date. This Skill automates the full sales workflow for Usell, handling both incoming messages and proactive re-engagement of stalled leads. ## Core Features & Use Cases - Reactive and proactive modes: Responds to incoming WhatsApp messages by detecting intent and pipeline stage, or initiates follow-ups with contacts inactive for over 24 hours. - CRM synchronization: Searches, creates, and updates contacts via MCP-CRM after every interaction, tracking pipeline_status, stage, ICP type, and timestamped notes. - Human-like delivery: Sends messages through the Evolution API with typing simulation delays calculated from message length, after passing text through a humanizer step. - Learning mechanisms: Escalates unknown objections to a Discord channel for owner validation, documents approved answers in OBJECTIONS.md, and commits updates to Git. - Use Case: A store owner replies "tá caro" to a pitch. The Skill matches the price objection in OBJECTIONS.md, applies the validated ROI reframe, updates the contact's pipeline stage in the CRM, and sends a natural three-sentence reply with a simulated typing delay. ## Quick Start Ask the agent to handle an incoming WhatsApp message from a lead by loading their CRM context, detecting the intent, and replying with the appropriate sales technique.
