Ready to deploy RPA? Avoid these four pitfalls

Sutherland Editorial

On paper, Robotic Process Automation seems easy, almost magical – Simply set the rules, deploy the robots, and enjoy the return on investment. Mundane tasks now get done with fewer humans, 100 percent accuracy, and abundant cost savings.

But there’s no sorcery involved with RPA. It’s rules-based automation that demands correct programming, planning, and stability, which means even the smallest mistakes in programming can have the biggest consequences. And that’s bad return on investment.

For example, automating inefficiencies or bad data will create problems down the road. Imagine a retailer that deploys RPA, but a SKU number in system has one incorrect digit and the customer orders and receives the wrong product due to the error. The retailer now has an unhappy customer and must pay for a return, causing a depreciation in the cost of that inventory. Now imagine this error repeated thousands of times by a robot, because its actions are only as good as its programming. The costs add up quickly.

To avoid mistakes in an RPA deployment, it’s a good practice to be aware of the traps that lurk in the shadows. Navigating through RPA deployment is much easier when companies know what to expect in terms of employee impact, costs, and logistics.

Here are four common pitfalls companies should avoid in an RPA project:

  1. Deploying RPA during a time of change
    If there’s one thing robots don’t like, it’s change. Robots prefer process maturity, consistent rules, and system stability. If a company makes changes too frequently to its back-office systems, the robot will struggle to keep up. The robot could even break due to constantly changing rules and upgrades. Robots can only do what they are rule-bound to do and machine learning can only reconcile certain system changes. RPA is most successful when deployed into a stable system.
  2. Deploying RPA into an unprepared organization
    Before making the investment in RPA, organizations must make sure their employees support the project and understand the risks they’ll face. An RPA deployment should not be a surprise to employees. They need reassurance their jobs aren’t threatened by the technology. In some cases, operations teams, particularly those not incentivized properly, will find any and every way to undermine the automation because they fear the robots will replace them. RPA is most successful when its adopters truly embrace the technology and are prepared for the changes automation brings.
  3. Deploying RPA into an organization with no plan for system maintenance
    Companies cannot expect to code a robot and then just walk away forever. Once RPA is up and running, the robots must be maintained and updated as a business changes and grows, or as its systems change. After new integrations, the robots will have to be introduced to those rules changes. Therefore, organizations need to have robust change management in place. They need to have good interaction between IT and operations to make sure they're in sync with those changes and confirm they have the right expectations so that the robots can react and adapt to the changes. RPA is most successful when its adopters plan for the financial costs to maintain and upgrade RPA systems and develop training programs necessary to keep employees up to speed.
  4. Deploying RPA too broadly
    There’s an old African proverb: “How do you eat an elephant? One bite at a time.” This holds true for RPA deployment. It is best to begin RPA with the easier processes, then add and build upon them, bringing in additional complexity and rules governances only after the foundation is strong. Don’t seek to automate the entire process all at once. Rather, break it up into smaller pieces and then add the different robotic processes together, integrate them, and then have the end-to-end solution. RPA is most successful when the implementation philosophy is crawl, walk, run.