Synopsis
- Client faced a lengthy and complex compliance testing process.
- Implemented Intelligent Automation, along with a cognitive decision service, to automate the testing process and classify client accounts.
- The new automated process saves a significant amount of time and enhances accuracy and reporting.
- As the account classification model gains more exposure, the system's accuracy is improved.
Full Article
In the world of real estate investment trusts (REITs) – companies that own and often operate various income-producing real estates like offices, apartments, and shopping centers – there's a critical need for regular compliance testing. Recognizing this, a prominent advisory, tax, and assurance consulting firm reached out to BP3 for an innovative solution to automate their REIT testing process. This process, legally mandated to be performed annually, is commonly conducted quarterly by many trusts to ensure continual compliance.
The firm's traditional testing method involved the cumbersome task of consolidating data from over 150 individual trial balance spreadsheets into a single, combined Excel workbook. Once consolidated, the data is categorized and updated to provide a view of all data that has been captured and manipulated by the automation process.
BP3's approach to this challenge was the implementation of Intelligent Automation, a combination of Robotics and Artificial Intelligence, to populate the firm's REIT testing Excel workbook with calculations, tables, and results reporting.
A key aspect of this automation was the creation of consolidated financial statements from various LLCs, each with its unique chart of accounts (COAs). The challenge arose when the robotic automation encountered COAs with differing account descriptions, necessitating a standardized mapping for the consolidated financials. To address this, BP3 implemented a cognitive decision service that activates when an unmapped account description is encountered. The decision service uses a combination of Natural Language Processing (NLP) and a Convolutional Neural Network to accurately predict the most likely mappings for these descriptions.
As the advisory firm continues to grow its REIT practice, the account classification model gains exposure to a wider range of COAs. Through ongoing training on these new accounts, the system's accuracy is improved and the robots become smarter.
BP3's solution not only streamlined the REIT testing process but also ensured that the firm could adapt and scale alongside its expanding practice, exemplifying our commitment to transforming complex business challenges into innovative, efficient solutions.