Just a few years ago, conversations about robotic process automation (RPA) were scattered and far between. Those familiar with RPA had a limited understanding of its potential and at best had carried out only a few isolated “proof of concept” experiments. Today, the RPA market is estimated to reach $2.9 billion by 2021, and conversations are becoming more frequent.

As a Forrester advisor on the Technology Council, I’ve partnered with senior technology leaders to find solutions to their challenges, identify opportunities to strategically leverage new technology, and connect them to research experts and their peers for support. While understanding of RPA has come a long way, there is still much to learn. As RPA moves from handling automation of simple processes to processes led by AI, it will be used for text analytics, chatbots, IoT sensors, and RPA platform efficiency.

Robotic process automation allows for scalable automated processes that connect systems in ways that were not possible in the past. RPA works by automating the easy, repetitive processes and then sending the more complex tasks to an individual with the expertise to handle more complex scenarios. Over time, the machine is trained to handle more and more of the complex tasks, allowing employees to take on even more complex scenarios or different roles. One member put it this way: “It’s like connecting systems through the front end. It’s like I have a robot sitting at the keyboard clicking away and occasionally being replaced by a robot that has better training.”

Today, I hear Technology Council members think about how to govern a robot. It needs a computer (usually a VM), it may need licenses for the software it uses, and it needs a way to take tasks in, mark them as done, or hand them off to someone else when the task is beyond its abilities. It also needs a manager in the business and coworkers who can check its work and identify when it needs more training. RPA providers are maturing, but current offerings and strategic maturities vary.

I asked an industry leader about his experience and how his company felt about RPA after having successfully stood up several robots and matured their execution of process over a few years. He is responsible for all application development and delivery, as well as being responsible for launching his firm’s EA practice.

“It’s been great. We can now process through these transactions much faster, and even the most complex scenarios have been incorporated into the bot. The hard lesson I learned is that once that knowledge is built into the bot and the employee goes out the door, it’s gone forever,” he said, seemingly exposing a risk that he hadn’t anticipated.

“What do you mean?” I said, thinking of my time documenting process as a business analyst. “Didn’t you record why the person was following those processes?”

“Well, we captured the process in the code, right?” he said, alluding to the idea that the code is the documentation, and then with a wink and a hint of sarcasm, he added, “I mean, the bot knows how to follow the process; we’ve just lost the business logic behind it,” countering the idea that the code itself is sufficient documentation.

“What would you do differently if you could do it again?” I said, thinking about how losing business logic and creating that kind of debt is happening with even the newest technologies.

“I’d still build the bot, and I’d document the business logic where I can. Most importantly, I’d find a way to keep the best employees whose roles are being replaced so their deep understanding of the business logic can be available as we continue to support our businesses. I mean, the business function that that system is used for is not going away, and having employees who have a deep understanding of our business is the hardest thing to hire for. We can teach them new skills and benefit from their expertise about us.”

It turns out the machine may replace the person doing the task, but a human is still needed to keep the understanding of the business. In this way, the robot may be a great partner, but it’s not ready to be an individual contributor.

Ready to execute your strategy successfully? Get tips and insights from a network of peers and analysts. Join Forrester’s Technology Council.