Company knowledge assistant: answers from your docs with cited sections, and routes the question to a human expert when no good source is found.

I build and hand over this system myself; the demo runs on sample data. Price & Timeline: $1,299 fixed | 5 business days | 3 revision rounds Tech Stack: Python 3.11, Claude AI, FastAPI, Vector Embeddings & Cosine Search, PII masking, Slack expert queue, Docker You provide: the Anthropic API key, hosting, Slack workspace access and an embeddings provider (or I use a local embedding model); their usage fees are on your side. The 5 business days start once I have access and a sample of your documents.

Demo Video (sample data): Architecture walkthrough (demo): https://youtu.be/Z6fRRQd-ELg

What you receive • Multi-format document parser (PDF/DOCX/MD) • PII masking step • Cosine similarity vector engine • Claude answers with cited sources, plus escalation to a subject-matter expert in Slack when no source is found

How it works

Kickoff & Access: Agree on the document set, the questions it should answer and who handles escalations; collect API access.

Architecture & Pipeline Build: Build the Python workflow with PII masking and human approval steps.

Integration & Testing: Connect your document sources and the Slack expert channel, then test with real questions, including ones your documents can't answer.

Handoff & Deployment: Deliver source code, Docker setup, setup guide and a recorded screen walkthrough with captions.