AI and Taxes

Reading the rest to determine how this is done.  Via determining algorithms via machine learning.  Note also the insertion of 'jurisdictional rules'

Deloitte uses AI to transform indirect tax recovery

Tax technology spotlight: CognitiveTax Insight™ solution uses machine learning and analytics to transform the tax office, saving time and reducing overpayments

NEW YORK, Aug. 8, 2019 /PRNewswire/ -- For decades, tax professionals have managed the analysis of complex data for indirect tax recovery through time-consuming statistical analysis and sample sets. Due to the sheer amount of data and complex jurisdictional rules, it was nearly impossible for many companies to analyze all indirect tax transactions. As a result, significant indirect tax refunds may not have been claimed. However, technological breakthroughs of the last few years are making this a challenge of the past. Deloitte has deployed a technology-enabled solution — CognitiveTax Insight™ (CogTax) — that provides an increased, improved, and more efficient analysis of clients' indirect tax data set. CogTax has the ability to analyze the full population, if desired, of a clients' accounts payable transactions compared to a traditional, sampled approach.

The solution not only allows tax professionals to save time and resources but can help companies to proactively avoid overpaying their indirect tax liabilities. CogTax leverages optical character recognition (OCR), along with advanced machine learning algorithms and analytics, to effectively analyze a full population of data and documents to assist clients with indirect tax overpayment recovery and reduce the potential for future over or underpayments.  ... "