Robotics Accounting Artificial Intelligence Tax Law Robotics Process Automation

Published: Mar 23, 2018
Topics: Tax
Share

Just as the new tax law spurs C-suite leaders in accounting, tax, and finance to reexamine their tax positions, the old ways of inputting and validating data no longer suffice. Automation is increasingly viewed as an efficient means of accounting for cash flow, repatriation income, and new state tax liabilities.

The need for real-time, predictive modeling that automation affords will only grow as the Supreme Court hears South Dakota v. Wayfair, Inc. later this summer and reconsiders current sales tax collection from online vendors, who are already seeing increased state audit scrutiny.

These common mistakes prevent successful automation across enterprise functions – in tax and finance, supply chain, operations, and human resources.

Fear of putting personnel out of work.

Let’s start with the elephant in the room. No one wants to be the one who puts somebody else out of a job. The truth is, robotics is coming, one way or another. But that doesn’t mean your personnel have to be left out of a job. “You can find other things for your people to do – instead of doing all data input, employees can raise their level of performance and become analysts, doing less repetitive work and tasks that involve more thinking,” said Gary Bingel, partner-in-charge of EisnerAmper's state and local tax group.

Overlooking current systems.

On the flip side, it may be tempting to jump into automation, but it’s equally imperative to thoroughly evaluate current systems first. Simply leading with technology in the hopes of finding a solution isn’t going to cut it. “Automation is kind of phase two,” Bingel said. “If your underlying data isn’t any good, you can automate all you want, but it’s like automating a driverless car that doesn’t run.” So first things first: Take a hard look at current procedures, and then layer in automation.

Only talking about potential solutions with vendors.

Automation is a journey. Simply having a vendor demonstrate a tool is insufficient. “You need to talk to your peers,” Bingel said. “Reach out to someone in your industry who has already started this journey and may be further along – see what they have done, because people have real-world experiences for the specific use cases that you’re looking for.”

Lacking a culture of innovation.

Understanding technology is one thing, but moving beyond the silo approach that typically plagues any successful implementation is the harder part. “Without leadership and sponsorship, these projects will have a difficult time getting traction,” said Gregory Fritsky, national director of robotics, AI, and data analytics. Making sure you have the right governance and program framework are essential, Fritsky said. In practical terms, leadership must set the vision; the program’s steering committee puts the governance model in place to support the goals and objectives.

Failure to pick the right use cases.

In assessing the effectiveness of automation on enterprise functions, some businesses pick use cases that are too small, with a negligible return on investment. A business may automate the capture of invoices, for instance, but stop there and, miss out on an opportunity to automate 70-100% of the end-to-end process. A business may also select a challenging process use case with complex rules and non-structured data, which is a bit complicated for robotics but better suited for machine learning or artificial intelligence solutions.

By contrast, repetitive and mundane finance processes, such as posting journal entries and performing reconciliations, can be easily automated using technology like robotics process automation. In such cases, automation should not merely be viewed as a cost exercise but a strategic one as well, allowing enterprises the ability to scale – and, in turn, better realize a solid return on investment.

Once you’re ready to tap an automation solution, remember: Artificial intelligence is a learning mechanism. “It’s taking in data and processing it – when it gets enough information, it can start to recognize patterns and invoke rules,” Fritsky said. “It needs time and investment – and a road map to get there.”

Business Tax Quarterly - Spring 2018