A.I. Needed to Decipher Tax Code

The 2017 Tax Cuts and Jobs Act included numerous pro-growth provisions, but it did little to simplify the tax code. Business taxation is particularly complex, partly because of the many tax credits that politicians have embedded to incentivize favored activities.

The research and development (R&D) tax credit is a good example. You may think that it is straightforward for businesses to add up the total they spend on R&D and multiple by a factor to find their credit amount. But the credit is so complicated that the largest and most sophisticated corporations hire specialists at the major accounting firms to calculate their claims. This guide to R&D tax credits is more than 700 pages long.

I received the following advertisement in my inbox today from a public relations company. Apparently, corporations contract with KPMG to calculate their R&D tax credits, then KPMG subcontracts with artificial intelligence experts at IBM Watson to help them out.

I wanted to get in touch today to share a recent success story from KPMG Research Credit Services. This particular division of KPMG works with IBM Watson to infuse the specialized knowledge of KPMG tax professionals with artificial intelligence capabilities with IBM’s artificial intelligence technology. This approach automates much of the qualitative analysis required to support an R&D tax credit claim by reviewing a variety of structured and unstructured data (i.e., R&D documentation) and then comparing the documents to relevant tax rules. The end result is more thorough and higher-quality documentation for a company’s IRS filings. And it saves time on research so that a company’s R&D professionals minimize their time on tax compliance activities.

KPMG’s tax professionals are feeding Watson with information so that it can learn to correctly and consistently pinpoint the right data that can help them assess a company’s qualifying R&D – often saving them cash along the way. Using intelligent automation in this way is a dramatically new approach to problem-solving for tax departments, which in the past have had to muscle their way through such challenges by adding staff. It’s also just one example of the way KPMG is taking advantage of intelligent automation, both internally to improve our own processes and externally to help their clients sharpen theirs.