Follow-up to the FOI Modernization Project – Ministry of Citizens’ Services
This case outlines the project for modernizing Freedom of Information Requests (FOI) for the British Columbia Ministry of Citizen Services. This project was successfully rolled out to 24 ministry clients by January 2024, and legacy FOI data( > 6TB ) was migrated. The next challenge was addressing one of the most labour-intensive stages of FOI processing – document redaction.
Over 2.5 million pages of documents are analyzed annually, from formats ranging from spreadsheets and PDFs to scanned handwritten notes and photographs.
The challenges were twofold:
- Volume & Complexity – More than 10,000 FOI requests per year, generating 600GB to 1TB of documents annually.
- Accuracy & Compliance – Each document must be carefully reviewed for Personally Identifiable Information (PII) before release, with a typical review costing the government $3,000 per request.
Manual redaction was time-consuming, costly, and prone to bottlenecks. The ministry required a solution that could balance speed, accuracy, and scalability while being compliant with privacy laws.
The Solution
In addition to the foundation of the cloud-native, microservices-enabled FOI Modernization platform, the AOT team incorporated AI-powered PII detection and redaction features based on Azure AI services.
Key features included:
- Custom AI Models for PII – Designed to identify sensitive data types like names, addresses, identification numbers, and other personal information in various formats.
- Advanced OCR & Search – Text extraction from scanned documents, images, and legacy archives for automated processing.
- Automated Document Workflow – Efficient ingestion, classification, deduplication, and redaction operations in the FOI system.
- Scalability & Flexibility – Handling a wide variety of file formats, such as .xlsx, .docx, .jpg, .png, .ppt, .msg, .ical, and .pdf.
- Business Analytics Integration – Offering visibility into processing times, redaction patterns, and performance gains.
Figure 1: High-level technical workflow for AI-driven PII detection and redaction in FOI processing.
This initiative was built on AOT’s SLED360 methodology to ensure compliance, speed, and alignment with public sector best practices.
See the Solution in Action
The demonstration video below showcases how FOI requests progress from submission to document review, and how Azure AI’s Document Intelligence and NLP Language Service automatically detect and redline PII across diverse document types, including scanned archives, handwritten notes, images, and PDFs.
Figure 2: Hybrid cloud architecture integrating Azure AI for scalable PII detection and document redaction.
The Impact
Introduction of AI-powered PII detection:
- Shortened processing times by as much as 6–7 days for FOI requests from pre-modernization timelines.
- Reduced manual burden for analysts handling thousands of pages per day.
- Increased accuracy in detecting sensitive information, minimizing the risk of overlooked redactions.
- Better enforcement of privacy and access legislation, facilitating safer and faster releases of information.
Looking Ahead
With AI-powered redaction in operation, the Ministry is looking into further capabilities to enhance FOI processes.
- Tailor-made NLP AI models for redlining non-PII material.
- Smart request redirection to third-party public authorities.
- Generative AI-fueled support robot for FOI analysts.
Through the automation of one of the most difficult steps in FOI processing and ongoing innovation, the FOI Modernization program is establishing a new benchmark for efficiency, compliance, and user experience in public sector information access. Backed by strong credentials in AI solutions and enterprise information systems, AOT Technologies is helping organizations reimagine what’s possible with modern, intelligent systems. Learn more about that here.