SalesBot

Sales reps at growing SaaS companies waste time hunting for information across Notion, Google Drive, and Slack. Responses to prospects take hours. New reps take weeks to get up to speed because the knowledge isn't accessible.

Building an AI assistant that sales reps query in natural language via Slack:

• Ingestion pipeline that indexes internal docs (product specs, pricing, competitor battle cards, objection handling guides, case studies) into a Qdrant vector database

• Claude API for retrieval-augmented generation: answers are grounded in the actual docs, with source citations

• n8n orchestrating the full pipeline: ingestion, chunking, embedding, retrieval, and response

• Slack integration so reps never leave their workflow

• Admin interface to add/update knowledge base documents

In development. Target: sub-5 second response time, sourced answers with doc references.

RAG-based assistants solve a real problem that wikis and CRMs don't: making institutional knowledge instantly accessible at the moment a rep needs it.