The first major category of automation we will dive into is Robotic Process Automation (RPA). RPA is not new. The core technology that RPA relies on (screen scraping) has actually been around for many years. One of the first RPA companies Blue Prism was founded 17 years ago and IPO'ed in 2016. RPA saw its first real customers during the economic downturn of 2008-2010, where most of the enterprises, especially financial services wanted to cut costs (VJ Anand). However, it was only in late 2016/early 2017 that RPA completely sky-rocketed.
One of the largest RPA players to ride this wave, UIPath saw its valuation reach 3B in 2018. In second, Automation Anywhere (AA) saw its valuation hit 2.6B post-money. This begs the first main question: why now?
A combination of business condition and technology maturity led to the rise of RPA. Nothing really earth shattering. RPA companies essentially started pouring large sums of money into Marketing campaigns that helped educate the market, resulting in RPA becoming an executive priority because of potential cost saving, optimization, and error reduction.
What about ROI? How do customers justify the 6-figure+ price tag for RPA? Initially companies evaluated RPA by calculating how many Full Time Employees (FTE) they could save per process automated. Recently, this has shifted more and more to how many FTE hours a process can save (Alex Qi). Essentially, this hard number boils down to: reduction of labor, reduction of error rates, and an increase in productivity.
Martin Roguljia, a Deloitte consultant, notes that after factoring in the cost of development, testing, toll gates from COE, a bot has to save more than 500 hours a year or 1/4 FTE to be justifiable. However, this number will go down as the cost of development decreases with friendlier and more stable technology. The market is headed towards rank and file employees building bots... however, we are very far from that right now.
What if we just let employees who know their processes the best create RPA bots? There are a few reasons why enterprises are scared to let their employees create their own RPA bots. Fears include:
Currently RPA is governed by a Center of Excellence (COE) within an enterprise. However there is a trend of moving from a central COE to a federated COE model where each department will have a COE. Each COE will get a budget they can spend on automation and will decide what comes in and out. This means that an organization can have multiple COE and each COE can purchase from different vendors.
On a technology side RPA vendors are building dashboards. A COE can use these dashboards to have a centralized view of all the bots that are running. This is a form of centralization of processes within an organization.