MCP Server
A custom MCP (Model Context Protocol) server built by Naidapa — compatible with prominent LLM models, it connects the AI to the company's system with structured access to items, suppliers and procurement data.
Case study
A Turkey-based commercial company whose procurement now runs through AI — Naidapa built an MCP server connecting prominent LLM models to their system, so the team names an item, the AI fetches suppliers, sends the RFQs, and reads every quotation email that comes back.

The company
Azure Commercial is a Turkey-based commercial company. Procurement is a core operation: the team constantly sources items from suppliers, requests quotations, and compares what comes back before buying.
Naidapa built Azure Commercial an MCP server — the layer that connects prominent LLM models directly to the company's system, so the AI works inside the procurement process instead of alongside it.
The challenge
Procurement eats hours in small steps: finding which suppliers carry an item, sending out requests for quotation one by one, then reading the replies — which arrive as plain-text emails, photographed documents and PDF attachments — and typing the prices and terms into the system by hand.
Every manual step is a delay and a chance for error: a quotation missed in an inbox, a price typed wrong, an RFQ that never went out. Azure Commercial needed the busywork automated — with the buying team still stating what they need and deciding what to buy.
The industry


What we implemented
The ERPNext modules Naidapa configured for Azure Commercial.
A custom MCP (Model Context Protocol) server built by Naidapa — compatible with prominent LLM models, it connects the AI to the company's system with structured access to items, suppliers and procurement data.
The team simply mentions the item they need to the AI. The AI searches the system and fetches the suppliers for that item — no manual digging through supplier records.
Requests for quotation are prepared and sent to those suppliers by the AI. The buying team states a need — the AI does the chasing.
Suppliers reply by email — as plain text, as an image, or as a PDF. The AI reads every format and extracts the prices, terms and details from each quotation.
The parsed quotation data is added straight into the system by the AI — no copy-typing from inboxes into records, and no quotation lost between the email and the system.
The outcome
Azure Commercial's procurement now runs through the AI: the team names an item, the AI fetches the suppliers from the system, sends the requests for quotation, reads every quotation email that returns — text, image or PDF — and enters the data into the system. The buying team starts from organized quotations in the system instead of a raw inbox.
Your project
Tell us how your business works today — we'll map it to ERPNext together.