KPMG has created a data analysis tool to extract relevant information from financial and other reports and plug it automatically into tax returns.
Tax Data Reader is designed to be able to read enormous amounts of data at once. The Big Four firm was awarded a U.S. patent for the tool based on its natural language processing and character recognition technology.
KPMG said the tool is especially useful for populating a Schedule K-1 (Form 1065), which typically includes a cover page followed by hundreds of pages of footnotes comprising unstructured data that needs to be sifted through and extracted for entry into a tax return.
“We’re using natural language processing and machine learning to convert a massive amount of unstructured data into structured data while saving hours of manual labor to input the information,” said Kevin Valek, KPMG tax partner and asset management tax leader, in a statement. “For one of our financial services clients, the tool works through more than 300 K-1 forms in less than 10 minutes, saving 80 percent in manual hours and resulting in a 50 percent increase in productivity on higher value tasks for their people. It’s a game changer.”
“With companies increasingly struggling with how to organize and make sense of their vast amounts of financial data, we’ve created a tool that eradicates their pain points,” added Brad Brown, KPMG partner and U.S. and global tax technology leader, also in a statement. “It’s all about making tax simpler via technology and equipping our clients with a sound data strategy so they can focus instead on running their businesses.”
The development of the Tax Data Reader is part of KPMG’s broader
