Whilst there are few organisations in Africa to have embarked on the Robotic Process Automation journey, Deloitte are predicting 'near universal' adoption by 2023. If you're a medium or large organisation, with ambitions to be competitive within your sector, chances are you will be on the RPA bandwagon.
So what is it, and what are the business benefits, and how do you get going?
RPA is using machines (bots) to replicate repetitive tasks which are currently being performed by people, using exactly the same systems. Herein lies the key difference between RPA and EDI (Electronic Data Interchange). Where EDI is machines communicating over a systems interface, RPA is machines using what was designed to be a human interface.
RPA vs. EDI
EDI is much, much faster to execute, and much less error-prone. So why would we ever want or need RPA? The truth is we don't necessarily want RPA, but the downside of EDI is the implementation complexity. It generally requires two organisations to be on the same pages of multiple books at the same time.
- They need to be aligned as to the business need / benefit, to avoid protracted 'who pays for it' discussions.
- They need to have resource available at the same time. One party can't test end-to-end without the co-operation of the other.
- They need to agree on the operational aspects of a 'standard' flow, plus exceptions and errors.
- They need to agree the technical standards for two (normally different) systems to talk to each other. This is normally the fastest, easiest and least contentious of any of the requirements.
So, how does RPA get around this? The key difference is where possible it takes the other party completely out of the equation. By using their input document (a PDF Purchase Order for example) as they would normally provide it, and training a 'bot' to read the PDF, and login to the transactional system to create a matching Sales Order, very little of the existing process has to change. Modern RPA will have the ability to respond to exception and error conditions by involving a human - via an email for example.
Benefits largely fall into the two camps of cost-savings, and reduced error-rates. A robot will generally work faster than a human - typically up to 10 times faster. 24 hours a day. 7 days a week. The economics are compelling for organisations with high document processing counts for the same document types. Organisations with small document counts for a particular document type will enjoy lower benefits as the bot needs to be trained for each document type. Shared service centers processing thousands, tens of thousands or hundreds of thousands of documents are where the business benefits can be staggering.
Reduced error-rates are as a result of much lower transposition error rates. Humans reading a number from an input document, and writing into another document inevitably make mistakes. A properly trained robot will faithfully move data from A to B time and time again.
Putting RPA into Action
To get started with RPA:
- Firstly evaluate your document processing volumes and variation
- Identify candidate processes for a pilot
- Identify the resource (internal and external) to work with
- Evaluate the technical solution to drive the initiative
- Develop a short Cost Benefit Analysis to confirm the business benefits are likely to be delivered
- Assess the results, evolve the program and scale up
The Human Impact of RPA
One step I've deliberately left out above is assessing and programming for the human impact. Not because it's an HR problem to deal with, so we don't have to, but because it is so important it needs a section to itself.
There is understandably significant fear, uncertainty and doubt within back-office employees at the prospects of the robots taking their jobs. This is justified. The reality is that there will be a human impact, and to overlook it is counter-productive. What will happen when roles disappear or are reconfigured? For each organisation the answer will be different. The best solution for all is redeploying within growth areas of a business. This may require retraining. If job loss is unavoidable, it's important to be clear which jobs, which people, when. There is nothing worse than uncertainty. Allow people the comfort of knowing how they will personally be affected and they are much more likely to either support the initiative, or at least not actively resist it.
RPA as Creator
I'm a firm believer that in the long run there will be no change to the pattern established by technology adoption over thousands of years. Technology creates jobs in the long terms. Jobs we have currently not even imagined, delivering services we didn't know we even needed (or, more accurately, wanted).