AI-assisted data submission analysis system for the Mandatory Energy Efficiency Labelling Scheme (MEELS)

I&T Wish AI-assisted data submission analysis system for the Mandatory Energy Efficiency Labelling Scheme (MEELS)
(REF: W-0605)
Matched I&T Solution
Trial Project
Summary and Challenges In this project, we aim to develop a comprehensive AI engine to automate the data entry and vetting processes for MEELS. The engine shall use AI assisted Optical Character Recognition (OCR) to interpret information (i.e. application form, laboratory test reports, Business Registration Certificate and supplementary documents). The system will automatically detect company chops and signatures on scanned paper submissions. The engine shall interface with existing EMSD databases for real-time data transfer and cross-checking the certificate validity. Furthermore, the AI system shall implement a self-learning mechanism to fine-tune model accuracy based on manual corrections, and provide an auto-drafting module for summary reports and correspondences.

Project Challenges:
  1. Different formats of test reports and certificates;
  2. Handwritten data in English and Chinese in application forms;
  3. Signature or company chop detection may be affected by the quality of paper scan;
  4. Limited availability of historical datasets for model training
Expected Outcome
  1. The Trained AI model(s) shall be capable of interpreting and extracting data from diverse documents including laboratory report formats and handwritten forms;
  2. The modular OCR engine shall be capable of multi-process document handling and certificate validity verification with automated user prompting;
  3. Automated detection and verification for company chops and signatures, including a cross-page consistency check mechanism shall be implemented to ensure regulatory compliance;
  4. A web-based interface for vetting workflow and a dashboard system for EMSD administrative analysis and reporting;
  5. An overall accuracy shall be 95% or above is expected. Detail accuracy requirements could be further defined according to scenario such as i. Printed text or Handwritten; ii. Different types of documents; iii. Document scanning qualities Some critical parameters may require higher accuracy (over 98%) which shall minimize manual intervention in the process;
  6. The system shall be able to interface with existing EMSD systems/databases for obtaining data and data comparison.
Expected Trial Duration 9-month
Contact Information
I&T Wish Proposer:Electrical and Mechanical Services Department (EMSD)
Contact Person:Mr. LAU Kwok Chi, Tommy
Position:Assistant IT Manager
Tel:3741 8868
Email: laukwokchi@emsd.gov.hk
Upload Date 2026-05-27
Closing Date 2026-06-10
Propose I&T Solution Submit Proposal