Robotic Process Automation and Intelligent Enterprise

From mainframe era to today’s Robotic Process Automation (RPA), new technology waves keep coming up every few years. Few years back it was ERP, the Internet, e-commerce, digital technology, Artificial Intelligence (AI), and now Robotic Process Automation (RPA).

Automation of manual tasks started much before the arrival of technology that we know today. IT is a subset of a broader technology landscape, and RPA is a subset of IT. So, when RPA is such a small subset of automation, why is it becoming a synonym for automation today? Why is RPA gaining so much attention and adoption suddenly?

What is RPA?

I came across this blog post that gives a comprehensive tutorial on RPA.

Just like its predecessors, RPA is also another piece of software automating manual tasks, left over by preceding technologies, especially ERP/SCM/CRM applications and Business Process Management (BPM) technologies.

From the functionality perspective an RPA bot is no different from copybook/sub-routine/function(mainframe), message (business integration), service (SOA), method (object-oriented) or upcoming micro-services. They all are performing a narrowly defined task that is repetitive.

However, RPA technology has been positioned little different from other information technologies, by giving a human touch to the piece of software in the name of a bot. A bot is positioned in parallel to human being, with its own identity.

I attended Automation Anywhere’s conference recently, where I saw a company reporting its work force count in terms of employee count, digital workforce count (number of bots), and total work force count. No wonder that Bill Gates proposed taxing robots similar to humans paying income tax.

What is the need for RPA technology?

As can be seen in the blog post (possible activities of RPA bots), activities performed by these RPA bots are mundane, clerical, repetitive and high volume, especially in large enterprises. A majority of these jobs are the result of mergers and acquisitions, distributed decision making across divisions on IT investments, budget constraints to replace legacy systems with new age technologies leading to co-existence of many different systems/applications in silos.

Some of these jobs were created by erstwhile ERP/CRP implementations, whose product vendors did not feel these tasks important enough to be automated in their products, and workflows designed in these applications is so deep that many end users are required to operate these applications (ERP vendors charge their customers based on number of end users, so there is no incentive for them to simplify work flows).

There have been attempts to connect the siloed systems/applications using workflow and BPM products, but they are risky and high cost solutions with limited success. These solutions require modifications to underlying systems/applications, which is risky.

Why have RPA products become so successful

An economic reason for their success is that bots work 24x7, whereas humans work in eight-hour shifts. Hence, at the same productivity level one bot can replace three humans. In reality a bot’s productivity is better than human productivity, theoretically one bot can replace four to five full-time employees (FTEs).There is no risk of attrition with bots, which is a serious problem with FTEs.

However, due to licensing policies limiting their use to specific department/desktops, sequential processing (unlike function or method that can be scheduled to run in multiple parallel processes, bots are prohibited from parallel processing) and task dependencies (next task may not be ready to be run, when bot completes previous task), these bots do have some idle time. RPA technology does not require much programming skills, and most of the work (creation of bots, scheduling them to perform tasks etc.) can be done by current end users themselves. So, talent availability, learning curve are no big constraints.

Sustenance of RPA technology

Since RPA products are addressing the limitations of existing systems, when those systems/applications get upgraded to fix these limitations, RPA bots may have to be retired. This is another reason why bots are comparable with humans, who also have a retirement age.

Does it mean that RPA technology vendors are short-lived? Those who have been in the industry for long enough, know how every time a new technology comes in, old technologies are written off on paper, especially by analyst community. In reality, we still have mainframes and IBM AS/400 machines running in some parts of the world. So, the limitations of existing applications may not go away any time soon.

However, just like ERP/CRP vendors growth plateaued after few years. The same thing can happen to RPA vendors.

The road ahead for RPA vendors

RPA vendors do understand this risk, so they have already started working on alternate technologies based on AI. They call it as Cognitive Process Automation (CPA), which is currently focused on addressing mundane tasks much like RPA, but deal with unstructured data such as images, text and in future could include voice, video also.

As AI technologies evolve, CPM products may become the new incarnation of erstwhile BPM products. This could lead to processes becoming context aware and intelligent, much like Google Search and Google smart reply/compose in Gmail. These capabilities at each business process level, could lead to creation of the Intelligent Enterprise, replacing the current transactional enterprise.

Development of processes and tools that enable an Intelligent Enterprise is where our startup company krtrimaɪQ is focused. To know more about Intelligent Enterprise and how to create one, please visit www.krtrimaiq.ai.