Why MCP matters for construction
The Model Context Protocol is becoming the standard way AI agents talk to external tools. Here's why that matters for an industry that still runs on phone calls and PDFs.
The construction industry is one of the last holdouts where "data integration" still means someone manually copying numbers from a supplier website into a spreadsheet. It works, but it doesn't scale, and it's exactly the kind of work that AI agents are built to eliminate.
The problem with supplier data
Australian trade suppliers — Bunnings, Nuway, Reece, and hundreds more — publish product data on their websites. But that data is designed for human eyes, not machine consumption. There's no API. No structured feed. Just HTML pages that change layout every few years.
For a human, this is fine. You search, you click, you read the specs. For an AI agent trying to source materials for a job, it's a wall. The agent can't reliably extract structured data from a webpage designed for browsing.
What MCP changes
The Model Context Protocol (MCP) is an open standard that defines how AI models interact with external tools and data sources. It uses JSON-RPC 2.0 over HTTP with SSE transport. It's already supported by Claude, ChatGPT, and a growing ecosystem of agent frameworks.
The key insight: MCP treats external data sources as *tools*, not as *content to be scraped*. An MCP server for Bunnings doesn't scrape the website — it provides structured tool calls that return structured data. The agent calls bunnings_search with a query, and gets back product names, SKUs, prices, and specs in a clean JSON format.
Why this matters for construction
Construction is a materials-heavy industry. Every job requires sourcing, pricing, and checking availability of hundreds of products. Today, that's done by humans with spreadsheets and phone calls. With MCP, an AI agent can:
1. Search for materials across multiple suppliers in seconds 2. Compare prices without visiting five different websites 3. Check stock levels by store location 4. Pull specifications for compliance checking 5. Build estimates by combining product data with labour calculations
The agent doesn't replace the human — it handles the data-heavy grunt work so the human can focus on decisions.
What we're building
TradeSDK is our implementation of this idea. We provide MCP servers for Bunnings Warehouse and Nuway Landscape Supplies, accessible through a single API key. The agent discovers available tools automatically via the MCP protocol and can start making calls immediately.
No scraping infrastructure to maintain. No stale CSVs. No custom integrations per supplier. One API key, standard protocol, real data.
The construction industry doesn't need another SaaS dashboard. It needs a data layer that agents can actually use. That's what MCP enables, and that's what we're building.