business objective
The client seeks Cognition’s assistance in gaining efficiency on a daily task by developing a solution that would:
- Efficiently extract M&A, fundraising, and IPO related information from SEC filings, including Form S-4, Form 8K, Form 10K, Form 10Q, Schedule 13D, Form S-1 and Form F-1, sourced from the SEC EDGAR database.
- The solution should eliminate the need for repeated keyword entries by leveraging contextual analysis and synonym matching to identify relevant M&A, fundraising, IPO, and strategic management change related information.
- Once the extracted material meets predefined relevance criteria, the program will generate a concise summary around the data.
Our solution
The Cognition team recommends an AI + Human Intelligence (HI) approach for this project, combining automated data extraction with human oversight for validation and refinement.
Step 1:
- The team developed a solution to extract information related to M&A, fundraising, IPOs, and management changes from SEC filings. A scheduler was implemented to automatically trigger the program at the desired frequency. To enhance accuracy, the program integrated synonym detection and contextual analysis, allowing it to identify relevant information even when acquisition-related keywords were not explicitly mentioned in the document.
Step 2:
- The extracted data was manually cross-verified against the original filings to assess its accuracy. Additionally, the original filings were thoroughly reviewed to ensure that no relevant information was overlooked by the automated solution.
Step 3:
- The solution was further trained and customized to improve data capture accuracy. Insights from manual validation were used to refine extraction logic, enhancing precision and reducing missed data points.
Outcome
- The turnaround time for identifying M&A, fundraising, and IPO-related financial data from SEC filings (with over 100 filings published daily) has reduced from two business days to approximately 8 hours.