I will build a custom ai agent with rag knowledge base for your business


Level 2
About this gig
Your team wastes hours answering the same questions from clients,
new hires, and internal staff that are already documented somewhere.
I build a custom AI agent that reads your actual business documents
(SOPs, PDFs, Notion pages, Google Docs, FAQs) and answers questions
instantly with accurate, source-grounded responses.
No hallucinations. No generic ChatGPT answers. Your knowledge. Your rules.
Here's what I build:
- Your documents ingested and indexed into a vector database
- AI reads and understands your content (PDFs, Docs, URLs, Notion)
- Smart retrieval pulls only the relevant context per question
- Deployed as a web chat widget, Slack bot, or WhatsApp assistant
- Fully hosted on your infrastructure you own everything
- New documents added anytime knowledge base stays up to date
Built with n8n + OpenAI/Claude + Supabase/Pinecone vector store.
- Works on any website, Slack workspace, or WhatsApp
- No hallucinated answers grounded in your real content
- Fully documented + video walkthrough included
- You own 100% of the system no vendor lock-in
Perfect for: agencies, SaaS teams, law firms, immigration consultancies,
HR departments, real estate teams, and client-facing support.
Get to know Yasir Shabbir
AI Automation Expert n8n Make Vibe Coding WordPress Lovable
Level 2
- FromPakistan
- Member sinceFeb 2020
- Last delivery2 weeks
Languages
Urdu, English
My Portfolio
FAQ
What file types can you ingest into the knowledge base?
A: PDFs, Word documents (.docx), plain text files, Google Docs (via API), Notion pages, website URLs, and CSV files. If you have a different format, message me before ordering.
Will the AI make up answers that aren't in my documents?
A: No. RAG systems retrieve relevant content from your documents first, then use that as the only context for the AI's response. If the answer isn't in your knowledge base, the agent says so rather than guessing.
What vector database do you use and why?
A: I use Supabase pgvector for most projects — it's free on the free tier, open source, and you already own the data. For larger-scale deployments (10,000+ documents), I recommend Pinecone. We'll decide together based on your volume.
Can I update the knowledge base myself after delivery?
A: Yes. Standard and Premium packages include an admin guide. Premium includes a full admin panel — drag and drop documents, no coding required.
Where does this get deployed?
A: On your own server or cloud account (Hostinger, DigitalOcean, AWS, or any VPS). You own 100% of the infrastructure. No dependency on me after delivery.
What does it cost to run after delivery?
A: OpenAI API costs roughly $5–20/month for typical business usage. Supabase free tier covers most small projects. No platform fees to me — you own the system.
Can this connect to my live website data or product catalog?
A: Yes — via URL ingestion or API-based data sync. This is covered in the Premium package's auto-sync pipeline. Mention it when ordering.
I'm an agency — can you build this for my clients?
A: Absolutely. I build on the client's infrastructure so they own everything. If you want white-label delivery (no mention of tools used), just request that before we start.
