Client background

This organization is responsible for monitoring the markets and financial services industry.

The business challenge

The company’s employee compliance reporting needed to monitor the market of public companies that were identified as a “conflict of interest” based on financial performance, ownership structure and industry type. The primary data source was comprised of self-reporting data with weak data governance and quality controls that had to be matched with financial performance, ownership data and other market monitoring on financial investments data. This included over 800 million permutations that had to be evaluated.

Strategy and solution

Baker Tilly worked with the company to implement an automated fuzzy matching process to recommend the best set of corporate entities and financial investments. The process utilized AWS SageMaker and natural language processing (NLP) approaches using edit distance and BERT embeddings to measure similarity with the corporate entities, financial investments and other market monitoring data.

This approach allowed the client to achieve the following:

• Reduced the hours to process the quarterly refresh by 35%
• Improved accuracy by picking up matches missed by a manual process
• Improved data quality and consistency to deliver by the reporting obligation dates