AI Case Studies: Regulatory Requirements

The Challenge

• 12,000 documents to process

• Estimated 205 days of manual work required – based on 10 minutes per document to match, locate, open and extract the information

• The information required was on the first two pages of the PDF files, time was saved using a standard read model combined with data extraction and pattern matching

• Tight deadlines incompatible with manual processing

The AI Solution

Standard Azure AI document processing model was used to quickly extract the information required. The solution utilized:

• Microsoft Azure Read model from Azure AI

• C# code to send the document to Azure AI endpoint and save the results into a SQL database

• Microsoft SQL code to extract key data for faster review

The Results

• Time Savings: Task completed in two days

• Efficiency: 99% reduction in processing time compared to manual methods

• Accuracy: High-quality data extraction with minimal errors

• Cost-Effectiveness: Significant reduction in labour costs

Key Benefits

• Scalability: Ability to handle large document volumes quickly

• Data Integrity: Inaccurate CRM data could lead to misinformed decision-making and flawed customer interactions

• Regulatory Compliance: Presenting incorrect information to a regulatory board could result in severe consequences, including financial penalties and even loss of licenses
  
• Accuracy: repetitive tasks performed manually are prone to human error, applying AI alleviated this risk

Conclusion

Our investigation revealed that the company’s Customer Relationship Management (CRM) system contained inaccurate data, potentially jeopardizing its reputation and invalidating its reporting.

The company could not only meet its regulatory obligations but also improve its overall data management practices and correct data in its CRM from the results.