Robotic Process Automation has been around for a while now, but there’s still a lot of misconception about what it actually is. Maybe you’ve heard different conflicting opinions and thoughts online about how it could impact you and how it can be used.
I want to highlight FIVE things that you need to know about RPA, its benefits and why you personally should get involved in automation. I also want you to be able to spot a process which eventually is going to get automated and give you a peak at popular RPA development platforms, so that you can see how easy it is to build automated “Bots”.
I’ve been a Lead Automation Expert for many, many years now, focusing on using “Lean thinking” with Intelligent Automation, working in many industries at some of the largest companies, tier one consultancies and also some small and medium enterprises. When I first heard about RPA was in 2017, when I was working as a Process Improvement Analyst (using lean six sigma to improve business processes). My boss had approached me to take lead on using “robotic automation” to improve processes and make savings. I honestly thought that he was about to roll out a physical robot from the storage cupboard behind him!
What is RPA?
To be clear, Robotic Process Automation is software robotics similar to an excel macro. RPA is low- or no-code where you can record the process you want to automate, or you can build the automated process from scratch by drag-and-dropping actions onto a process flow. RPA is able to scrape data from website pages or web-tables, or directly from databases using APIs. It can add data into an excel or fill in webforms or into desktop applications.
This highly versatile application can work with any application desktop, web app or website as it uses the User Interface just like you and I would. It can press buttons and links, tick boxes, select from dropdown menus and type into any text field as if it were using a mouse and keyboard.
It is an amazing technology but it’s not actually a silver bullet. There’s a lot of things that RPA can’t do and that’s why it businesses are coupling RPA with AI capabilities. Pure RPA can still automate potentially hundreds of your business’s processes right now, but RPA is limited by the fact that it can only use standardised input and can only make logical decisions. I.e. RPA can’t do anything which is which would require human judgment and intuition.
More and more RPA is overlapping with Artificial Intelligence (AI), as vendors add AI capabilities into their software so that the scope of what RPA can do for your business keeps increasing. Gartner had estimated that RPA on it’s own would be able to eliminate 20% of repetitive tasks. When I was interviewed for Process Excellence Network in 2019, I shared how this trend would eventually bring an end to pure RPA.
With the emergence of Intelligent Automation (RPA + AI), if say you wanted to read data off a pdf or respond to a chat, or use information collected by a chatbot to better serve a customer, you can now literally drag that capability into your RPA bots workflow.
The Optical Character recognition (OCR) software is able to read the PDF document (whether that be text or handwriting). A Natural Language Progressing (NLP) integrated tool could understand the customers intention from the question entered into a chatbot/webchat, and then trigger an automated process to take over and carry out the required action or retrieve requested information.
What are the benefits of automation?
Robotic process automation on it’s own is great at manually intensive tasks, that are tedious and repetitive, and require a lot of combined staff effort to manage the workload. RPA is especially good for those repetitive tasks that have seasonal volume.
Do you have areas in your business where staff are constantly working late?
These teams could leverage automation to save them time on odd-jobs and partially automate processes here and there to free up more of their time to hit their targets and deadlines faster.
Do you have KPIs showing high amounts of human error in certain teams?
Automating processes prone for error can effectively reduce errors to near zero. Even if the whole end-to-end process can’t be automated, augmenting staff can ensure they retrieve/input/use accurate information or data validation can be automated.
Do you have long physical queues or customers on hold for long period of time? Perhaps your staff are taking a long time to respond to clients.
Intelligent automation has grown very popular in the customer service division. Customer service advisors use multiple applications to assist enquiries and also have a lot of form filling and updating customer data so automation can be used to streamline a lot of these processes
With seasonal spikes, instead of going through the whole process of finding and hiring new teams and having to train them up, your existing staff can leverage automation technology to be more productive.
Are you outsourcing certain processes that could be automated?
You could potentially develop a Bot once and all you need to pay after that is server costs, the software licenses and maybe a little maintenance to keep it running and up to date
The by-product of these cost savings, improved efficiencies, and error reductions, is the enhancement of customer experience. The service that you deliver to customers will be more accurate, faster and more convenient that from your competitors that don’t use automation effectively.
Giving automation technology to your staff can also improve staff morale as it can be like giving everyone in your team a virtual assistant to exceed targets and gain a better work-life balance with fewer late nights and more time spend on interesting, creative, human to human interaction and customer facing tasks.
How do I spot an RPA use case?
There are about eight things you want to look for when identifying a process which is suitable for automation:
1. RPA needs standardized input. Its able to use pre-determined information like options in drop down menus or selectors. RPA on its own can’t understand free text
2. RPA can’t do a process that requires human intuition, RPA cannot think, it can’t make any judgment calls. It pretty much can just ‘copy’ a task if you show it the steps
3. RPA needs rules. It can only make a decision if its rules based, and the outcomes are pre-determined. You must be able to program the logic that the Bot needs to make to choice the right option.
a. For example: IF price goes up press “BUY”, ELSE IF price goes down press “SELL”
4. RPA works best when the process is repetitive and manually intensive. You want to find processes which repeat over and over again like updating details for a long list of customers
5. RPA is best for those high volume processes like adding 100 customer details every day into a CRM application. This isn’t a necessity, but it does mean that you will deliver a significant amount of savings for your development effort.
6. The process needs to be stable you if your RPA is to last. RPA using the User Interface (clicking on buttons, types in text boxes). If the process keeps changing the Bot will keep crashing and will constantly require new development so that it can navigate the new application or webpage layouts.
7. Make sure there’s no pending changes on the process or applications. Related to the previous point, it’s not a good idea to have your team working on automating a process that is due to change in under 6 months’ time, or the applications are due to be upgraded. This could require having to rebuild the automation before the end users ever had a change to properly benefit from it. Changes are inevitable, but timing is everything.
8. Obviously, the process for automation must be fully computer-based. However, it certain steps require someone to pick up a phone or print something then it could be possible to automate with a “human in the loop”, where you automate the first part of the process, the bot notifies the person to do a physical task, then that person triggers the automation to finish that process.
a. It is possible to integrate RPA (software robotics) with physical robotics, for example, an automated payment process triggers a physical robot to pick the stock in a warehouse
How can I build an Automated process?
The RPA studio is where developers design the automated steps for the Bot to follow. There are two main types; drag and drop where you can just drag an activity like an action or decision point into a process map (you would see this set up on Blue Prism or UiPath software).
The other way is displayed as line-by-line code, where you drag the activity and all steps show are lines of code in a page, instead of shapes (see the images for a better understanding). This is typicaly found with Automation Anywhere and IBM’s software, however recently there is an option to depict the process steps with shapes.
To build a robot there are just a handful of steps you need to follow to create a building an automation:
IDENTIFY PAGE ELEMENTS: you have to spy different objects on a screen or a page of a website or an application and all this is it just helps the rpa robot to recognize different elements on a page so this is a button this is a link this is the name field
DRAG ACTIVITIES/ACTIONS: you can drag in different activities into the process map like an action to open excel, open chrome or type a web address
DRAG DECISION POINTS: you can drag in a decision so that could be:
IF price had gone up press “BUY”
ELSE IF price has gone down press “SELL”
ADD LOOPS: this loops through the process for a predefined number of iterations. For example you could have a data table in excel. The bot counts the rows and loops through the table row by row inputting data from excel into a web form
PRESS PLAY: Once the process is finished you can press play (or define the trigger to start the process) and you will see the robot moving through the process step by step
That’s it! Pretty easy right?
RPA Vendors are quickly moving towards a world of “Citizen Development” as they develop their platforms to become so intuitive, anyone could build automated processes.
How is RPA using/integrating with Artificial Intelligence?
RPA on its own can probably only automate about 20 percent of processes (still you’re looking at hundreds of processes), but then what about a rest?
This is where you need different types of AI to plug into your automated process so that it can extend its capability. RPA platforms are making it really easy to do in two ways. Either they’re embedding AI into their platform or they’re building massive ecosystems where it’s very easy to integrate with different AI applications.
Let’s take a look at a few:
OCR (optical character recognition) is a type of AI that can read data from an image. So if you had a pdf or a scanned image or maybe even just a photo, Optical character recognition (or perhaps Intelligent Character Recognition) can turn data from an image into structured information that the RPA Bot can use. You could get a scanned invoice or PDF and with OCR capabilities your automation can input invoice data directing into your excel spreadsheet or your finance application. Intelligent character recognition (the smarter version of OCR) cans understand things like handwriting, varying fonts and even spelling mistakes.
Another very popular AI in the office is the chatbots. You can have a web chat where customers are asking the same questions like “what’s my balance?” or “who’s my account manager?”. Maybe customers wants to do similar activities like “I want to purchase flights to cyprus in january” or “Book a hotel” or “book a car”. A chatbot using natural language processing (NLP) to understand the intent of a question can trigger an automated process to go into the customer’s account and find who the account manager is, or look at their balance. If it were a holiday making site, the automation could book a holiday/flight/car using the requirements the customer specified.
NLP, however, is not limited to chatbots. It can understand free text, so could read an email or text in a comments box. For example, NLP capability can read the content of an email and use this to categorise and organise a mailbox, which in turn could trigger different automated processes for each type of email. Intelligent automation has endless possibilities.
Machine learning and this is probably what most people think is “AI”. You essentially feed a computer with a load of data, and it finds different patterns and correlations, so that when you ask it a relevant question it can calculate the answer which has the highest probability of being correct.
This has been useful in healthcare. The computer is given the details and symptoms of patients that have a specific disease, and the details of people that don’t have that illness. By calculating the similarities and dissimilarities between the people in the two groups, when the machine is given the details of a new person, the machine can calculate the probability that this new person as the disease or not.
Sales and Marketing can use machine learning in a different way. If you have hundreds or thousands of customers that you want to sell something to but you don’t fully understand their buying habits or preferences. By feeding customer data and purchase histories to the machine, it can calculate patterns can make the best recommendations of new products to customers that have a high probability of buying.
Now that you have an awareness of how Robotic process automation is being used in the office and how Artificial intelligent is supercharging what RPA can do, start looking at processes in your business that can be automated today
If you want to learn more about Intelligent Automation in your office, subscribe to my YouTube channel Tony IA (Intelligent Automation, Simplified) for videos created weekly, to simplify intelligent automation for business leaders and professionals who are new to automation to level-up your knowledge. Become empowered on how you optimise your business and discover new technologies, in a lean and accelerate way. You can also learn more from my book, Business @ the Speed of Bots: The AEIO YOU method HOW TO IMPLEMENT ROBOTIC PROCESS AUTOMATION THAT SCALES. Get ready for the new digital transformation age for more information. The foreword is written by Guy Kirkwood, who is the Chief Evangelist at UiPath, and a very well-known advocate of RPA with over 20 years of experience in outsourcing.