RPA: The Duck Tape of the IT World

I was really disappointed recently when I went to my first Robot Process Automation (RPA) talk at an insurance conference. From the excited buzz I heard about RPA, I was looking forward to seeing something new and different. I think my imagination probably led me to envision some sci-fi type robots at keyboards. After 10 minutes, I realized I wasn’t watching a sci-fi vision of the future rather a part of my past re-emerging.

In the 1990’s I worked on a project at an insurance company to implement a GUI interface for the customer services team who worked on IBM mainframe green screens. The new GUI interface ran on IBM PS/2 series computers on a token ring network with terminal emulation software and special scripting language to populate the GUI. In other words, screen scraping. The project went well. We learned a lot and after some productivity hits due to the change to the service reps keyboard changes, most were happy.

What I saw in the RPA session at the conference was a modern version of screen scraping, with better tools and hardware. There were multiple systems tied together through the RPA layer and some new capabilities implemented at that layer as well and again, people were happy with what they got. The insurer presenting had established a hefty RPA team within the IT department with its own COE, program management office and development team.

Since then, there have been several articles and case studies on using RPA with legacy insurance systems to preserve the life of existing systems and avoid the need to do “ruinous” transformation programs for many years. And it certainly can work for a time and resolve very specific problems. But in the end, it’s still Duck tape.

Duck tape is an amazing thing, used for 100s of repairs, creative lash-ups and unique solutions to small problems. It can hold together an object that has developed cracks and sags, and connect two objects that were never meant to be connected. And in many ways, RPA is the Duck tape of the modern era.

Just a few thoughts on Duck tape:

Duck tape works best on things that don’t move

Anyone who has Duck taped moving parts together knows eventually a combination of torque and Duck tape’s flexibility will lead to a tear in the tape. In RPA, there is an assumption that the legacy systems that are being automated don’t ever change. This is not always true for core insurance systems as regulatory or product changes may require changes to the core legacy system which in turn can break the automation through changed inputs/outputs. But that’s OK because there is always more Duck tape that can be applied.

Duck tape does nothing to enhance the capabilities of the item it wraps

RPA, like Duck tape, is a container and binder, not changing the nature of what it contains. Unlike Duck tape, RPA can add additional capabilities but they are additive to the core legacy system capabilities and can require a lot of work to implement around legacy architecture and process patterns. At some point, the complexity and cost of the RPA rivals building new modules that can replace core legacy capabilities more simply and with less fragility.

Duck tape will tear, and you must plan for that

Duck tape was developed to seal ammo boxes for soldiers in World War II until needed and then tear and be discarded. Duck tape was never meant to be a permanent solution. Whether it’s fixing a seam, taping a car bumper back onto your car or sealing a hole in your boat to make it back to shore, it’s temporary. RPA is more robust in this regard as under the right conditions it can be used in the longer term but there are always potential tears including:

  • The underlying legacy system becomes erratic due to unexpected system bugs that only occur after decades of transaction processing
  • The underlying legacy system become unsupportable due to programmer attrition (you still need a few to deal with the above)
  • The underlying legacy programs system environment becomes unavailable (think Dec VAX)
  • The underlying legacy systems system environment becomes prohibitively expensive (think mainframe system tools)
  • Modern security requirements can’t be met as the core legacy system interacts with the RPA tool
  • The CEO tells you everything must be in the cloud in 18 months

I keep a roll of Duck tape in my car and backpack at all times because I know I will need it, and I would use RPA when appropriate to make a system last a little longer or add some additional capability to an existing process to meet a looming deadline. But I would not mark that project as done until I planned for the Duck tape tearing. Hopefully, the  RPA implementation was buying time to do the core system replacement project that will provide the long-term solution.

If you’re worried about how much Duck tape is on your legacy system, I’d welcome the opportunity to chat with you about the FINEOS Platform and how it can enable your insurance business to ensure accurate, open and flexible product development.