AI/SaaS / Maintained
Voice-Driven CRM Assistant
Drop a voice note in Slack — the assistant transcribes it with Whisper, uses GPT-4o function calling to decide the right CRM action (create contact, log activity, query, or answer), runs it against Supabase, and replies with a spoken confirmation. Short-term memory keeps multi-turn context across requests.
Role
Designer and engineer (solo)
Disciplines
AI Integration / Backend Systems
Status
Maintained
System overview
Full conversational loop: Slack file_shared event → Whisper STT → GPT-4o with three strict-schema function definitions (create_contact, log_activity, query_contact) and tool_choice: auto → Switch node dispatches to the right Supabase action → spoken reply synthesized with ElevenLabs and returned as a Slack file. Short-term memory via Supabase: last 10 turns per user replayed into the message array for follow-up requests. Defensive handling for Slack URL verification, mime-type gating, Whisper hallucination detection, and error sub-workflow.
Validation and results
Working voice CRM loop with multi-turn memory, three LLM-dispatched function tools, production-grade error handling, and all the defensive edge cases handled — Slack handshake, hallucination detection, mime-type gating.
Media gallery
