Implementing RPA is easier than you think
AICPA logo
AICPA logo
  • Home
distorted colorful image of random numbers

Implementing RPA is easier than you think

3 years ago · 3 min read · AICPA Insights Blog

Robotic process automation (RPA) has advanced quickly in just a few years. In a short amount of time, both the number of RPA providers and the range of capabilities have greatly increased. So much so, it can be difficult to keep up with all the changes.

This leads to the misconception that RPA is either too complicated, too costly, too technical or simply too difficult, for small or medium-sized businesses to put in place. While these statements may have once been true, that is no longer the case. Today, the wide variety of available options mean RPA is now for everyone. The barriers to entry have dropped so now is the right time to reconsider how automation can help your business.

What’s RPA?

Ever seen an excel macro run? It looks like an invisible person is moving data around in your spreadsheet, reformatting cells, and summarising information. Macros are known to consolidate and reform data, but they have a major restriction: they can only manipulate data in your spreadsheet.

RPA is similar to the excel macro, but unlike the macro isn’t restricted to a single application and can be used on any desktop application you use on a day-to-day basis. Normal tasks that use a system can now be automated.

Copying files, renaming them, collating information, reformatting data, copying information… Repetitive tasks like these are game for RPA.

Learning to Understand

The moment you start to understand what RPA can do and overcome the misconceptions that initially excluded it from your thinking, the next challenge is finding the chance to apply it in your own business. You can find many demonstration videos showing RPA in action, but it can be difficult to implement it in your own office. There’s a lag between understanding what RPA can do and understanding how RPA can be used. Don’t worry, the opportunities are there.

Now you’re faced with two possibilities: seek help or jump right in. The first choice will require funding but will speed up the process, the second seems more daunting, but shouldn’t.

There are low-priced, modest, and effective options available to get you started: start small, think big, grow fast, as the mantra goes.

Because RPA acts as the glue to link isolated systems, it’s important to stress that it works just as well as a solution for integration as it does for automation. The Financial Services industry is a perfect example. Companies in this industry have often grown through acquisitions and end up with multiple, separate IT systems all with similar functionality. RPA can connect them and increase the data integrity between them. Even small businesses can have their own applications to handle sales, accounting and ordering resulting in manually re-entering key information into separate systems. The lack of integration isn’t due to time or acquisition. It’s due to the modular way system solutions are bought.

Let’s make automation simple: It’s repetitive transactions that uses structured data. As a result, speed, customer satisfaction and quality are all improved.

Up until recently, that was all RPA could do. Now, artificial intelligence (AI) is changing that.

Commoditizing AI

Mass media occasionally exaggerates things, which can lead people to misunderstand what AI is and what expectations are unattainable. Fortunately, the market is changing, and we’re witnessing a blending of RPA with Machine Learning (ML) to address more intricate transactions that were previously out of scope for RPA on its own.

The difference between machine learning and traditional programming is that ML learns by example, and traditional programming involves developers explicitly coding every situation. A clear example of this is character recognition. In character recognition, traditional coding requires every variation of handwriting to be known and coded in advance (which is highly unlikely), however, ML identifies characters after being offered just few hundred examples.

Initially, data scientists were scarce, making it hard for new businesses to use new ML capabilities. You can now buy ML algorithms off the shelf, meaning they’ve already been trained. For example, they can understand the information contained on an invoice scan and launch the necessary postings in accounting systems.

Together, RPA and ML can yield solutions to difficult business problems which makes it tempting to start there. Take it from me — start with the simple problems RPA can handle first: start small, think big, grow fast.

More about RPA

There is lots of free information to get you started with RPA.

For more, you can check out our "A-E" of digital disruption learning series starting with Automation.

Make sure you test your knowledge here by taking our automation quiz.

Rob King

Rob King is the author of Digital Workforce, an executive guide to RPA. It explains the different RPA vendors, along with elements such as roles and structures needed, key components of governance and processes for successful implementation.

What did you think of this?

Every bit of feedback you provide will help us improve your experience

What did you think of this?

Every bit of feedback you provide will help us improve your experience

Mentioned in this article



Manage preferences

Related content