Robotic Process Automation (RPA) bots are ignored both when they make a mistake and when they get it right. That’s leading to some major problems for businesses that rely on them

Many businesses have come to rely on Robotic Process Automation (RPA) ‘bots’ to automate repetitive tasks, but not all of those businesses have thought things through.

RPAs are essentially the modern version of ‘send keys’ where you can program the computer to hit certain keys in succession to execute a task over and over again. The difference is that RPAs are web enabled, can do more complex actions, and can use machine learning to do it better. They work great for high volume tasks that would normally take too much of an employee’s time. Popular RPAs include Microsoft’s Power Automate and UiPath.

Not as simple as ‘set it and forget it’

RPAs are specifically programmed to do a task the same way over and over, which is great, but that’s also where things can go wrong. What if even the slightest change happens to the environment in which the RPA is working in? That could mean someone adding a row to a spreadsheet, using a different field name, or even changes that happen automatically as part of a software update.

Sometimes a change won’t affect the RPA, but if it does it can throw the entire thing completely off. If things are off even a hair, the RPA could be ‘fetching’ the wrong piece of data or confirming whether or not work is complete based on the wrong field.

Worse, these mistakes are often missed for weeks before someone catches them. Sometimes it takes a client saying something like, “what happened to those financial reports my department was getting? It’s been 3 weeks since we’ve seen one, and I just found out that we’ve blown our budget” for a business to notice the RPA is silently failing.

Stop ignoring your invisible robots

RPAs need to be managed like an employee. They should be checked on to make sure they’re doing a ‘good job’. Otherwise, they will keep making the same mistakes over and over again.

But why do these mistakes often get ignored for so long? With typical RPAs, you can set ‘expectations’ on what result you will receive. Imagine something like a valid range. If something is outside of this valid range, the RPA automatically sends an ‘exception report’ to a pre-set user.
 
So, what’s the problem? The problem is that these exception reports are sent by email, usually to someone who manages the application rather than works in it. As we all know, emails are extremely easy to ignore, especially when we get 50-100 of them each day.

Think of your IT person’s email inbox. It probably contains endless help desk tickets, back and forth emails, and notifications from various apps. It’s extremely easy for an exception report to get lost in the noise, allowing the errors to keep repeating for weeks until someone notices them.

Even worse, in some cases the RPA thinks it has completed successfully and sends no error message at all. This can occur when pressing the tab key three times used to take you to a field with a check amount, and the RPA was to report the check amounts larger than $5,000, but now pressing the tab key three times takes you to the field that recorded the time the payment came in. Because that number will never be higher than $5,000 and there will be a number in the field, the RPA will never escalate the work or report an error. This is where you never find out about the problem until it grows big enough to get noticed by a human.

RPA “Control Towers” are not the answer

Many RPA vendors offer “Control Tower” technology that lets an IT professional monitor the activity of numerous RPA bots from a central location. These provide an excellent mechanism for discovering when an RPA has simply “crashed” or is signaling errors. This is significantly better than just sending out an email and hoping someone reads it, but it still assumes that the RPA bot is aware that it’s having a problem.

As we mentioned in the last example, if the bot thinks it’s doing just fine, so will the control tower. Additionally, the vendor control towers only support that same vendor’s RPA bots. In the real world, most enterprise organizations end up deploying multiple RPA technologies. That’s why you need a solution that can interact with bots that have “different personalities”.

HighGear lets you manage RPAs just like they were a human employee

At HighGear, we firmly believe that you cannot just turn on your RPA and then ignore it while it ‘works’. That’s why we’ve built tools to connect with and manage RPAs as if they were another one of your employees.

HighGear watches the RPA to make sure it’s doing the work correctly. RPA bots are assigned a task either through HighGear or through integration. HighGear then tracks the task’s status, and if a corrective action becomes necessary, HighGear will automatically route the task to a human or team and send them an alert.

Instead of just getting an exception report in an email, HighGear keeps a running list of exceptions that can be actively managed and routed by specialists. You can also watch exceptions over time to find patterns and common mistakes and build new ways of solving them.

One of the nice things about this is that it’s all color coded in an easy-to-use dashboard. The dashboard acts as a central hub to manage all your RPAs, regardless of which RPA vendor(s) you are working with. Finally, you can congratulate your digital workers  for the all the green ‘task completed successfully’ boxes and quickly take corrective action on all of the red ‘bot failed’ boxes.

Helping you treat your robots better

With HighGear, you’re guiding your robots to success, rather than just turning them on and ignoring them. After all, they’re doing important work, just like the rest of us.

You wouldn’t ignore a data entry intern or entry-level employee. You’d give them a job to do and coach them along the way. RPAs are doing some of the same work that those people did just a few years back, so why ignore them?