Project Proposal: AI-Driven Commissioning Excellence
A Retrieval-Augmented Generation (RAG) AI Agent built to transform the Commissioning Department from a memory-dependent workflow to a data-verified powerhouse.
Executive Summary
The Commissioning Department relies heavily on complex PDF documents, including the CIBSE Codes and Standards and ASHRAE Guides. The standard procedure for addressing MEP installation issues involves citing these standards to justify site instructions or non-compliance reports issued to contractors.
The Problem
Finding the exact reference within hundreds of pages is highly time-consuming. Engineers often depend on memory or manual searching (CTRL+F) through massive, unindexed PDF files. This inefficiency causes delays in issuing reports and weakens the firm's negotiating position during technical disputes.
The AI Solution: RAG Agent
A Retrieval-Augmented Generation (RAG) AI Agent built on the n8n platform. This intelligent workflow ingests industry codes (CIBSE, ASHRAE, local UAE mandates) into a vector database (Supabase) and allows engineers to query them using natural language via a Slack/Teams integration or a web interface.
Core Technology Stack
- Vector Database: Supabase
- Workflow Logic: n8n
- AI Model: Gemini 3.5 Flash Lite API Integration
Quantifiable Results
Beyond the data, the pilot study showed a significant improvement in Client Confidence. When engineers can present the exact section and reference number instantly, technical directives are respected immediately, reducing project friction.
Conclusion
The RAG AI Agent transforms the Commissioning Department from a memory-dependent workflow to a data-verified powerhouse. It bridges the gap between senior expertise and junior execution while providing the consultancy firm with a distinct competitive advantage in the UAE market.