Robotic Process Automation – The Digital Worker for an Intelligent Enterprise
Published on : Friday 02-09-2022
RPA generates business value by improving operations to mobilise resources for high-value tasks at lower costs, says Satyadeep Dey.
As organisations across industries strive to adapt quickly to the demands of a never changing and challenging business environment, create new compelling products and services, improve sales and customer retention, reduce cost, and increase employee productivity and engagement there is a need to streamline their existing business processes. Every business, large and small, follows some processes to achieve their business goals. These processes are usually complex and often span multiple applications from different vendors built using a variety of technology, both old and new. These applications could be deployed on-premise, in the cloud or in a hybrid environment. To add to the complexity, many steps in a process continue to be manual and are therefore prone to errors, inconsistencies, and inefficiencies.
Robotic Process Automation (RPA) is a critical new technology that helps organisations optimise processes by automating them.
Robotic Process Automation
Robotic Process Automation (RPA) accelerates digital transformation of business processes by automating repetitive and tedious actions where human workers don’t bring any added value. It frees up workers and gives them the bandwidth to work on higher value tasks.
Take the simple example of a data entry operator who reads data from some source, say Excel, and enters this into an ERP system. This task can be easily automated using an RPA bot.
RPA bots can be used in 2 modes –
i. As a Digital Assistant (or attended bot) where the bot works along with a human and helps the human perform some tasks, and
ii. As a Digital Worker (or un-attended bot) where the bot works independently with human intervention limited to checking status and handling exceptions.
RPA solutions automate repetitive tasks, traditionally performed by humans, without modifying the business systems. This capability of automating tasks without the need to change the underlying applications is a huge differentiator for RPA. Take an example where you have a process that spans two legacy systems and neither has APIs which can be used for integration. An RPA bot can be designed to interact with the UI of both these systems, like what a human user would do, and thus help in seamlessly moving data between these two systems. This gives businesses the ability to make quick tactical changes while more long-term strategic changes, like upgrading to more modern applications, can be done in parallel.
Most RPA tools are designed to be used by citizen developers and increasingly RPA tools are adapting to the Low-Code-No-Code paradigm with ease of use being a primary consideration. You don’t need to be an expert programmer to use these tools which means that more people in an organisation, including business users, can build RPA bots. This is a huge advantage from a cost and capacity perspective.
Many technologies complement and work along with RPA and these days it’s quite common for these complementary capabilities to be bundled together by RPA vendors. Artificial Intelligence, Machine Learning (AI/ML) is one such technology. Machines have always been better than humans at processing structured data – performing mathematical calculations is probably the best example. Advances in Computer Vision and Natural Language Processing means that computers can now also handle unstructured data. This opens up a world of business opportunities – ML models can identify and classify multiple objects in an image, do machine translations, voice recognition, recommend actions, read scanned and hand-written documents, correlate and contextualise information and perform many other tasks that could only be done by humans a decade back.
Let’s now look at some real use cases from industry and how RPA has helped solve some tricky problems.
Automated invoice entry into ERP system
A pharmaceuticals major had a challenging situation with their invoicing process during the Covid-19 crisis.
Invoice receipt and payment was a process managed by employees who received e-mails with scanned invoices attached. They then manually created invoices in an ERP system. The process was slow, with each invoice taking several minutes to enter and it was also error prone and often required additional effort for correcting wrong data-entries.
Covid made the situation even more challenging. There was a need to scale up the operations so that it could function 27x7 with minimal human intervention and account for unplanned leaves of employees due to illness.
The process flow is as shown in Figure 2.
1. Invoices are received as attachment in e-mails
2. The RPA BOT extracts the invoice
3. Invoice is sent to a document processing service which is powered by Machine Learning
4. This service can read and contextualise the information, i.e., map a value to a label, and
5. This data is then entered into an ERP system to create an invoice.
In this use case the RPA BOT behaves like a human worker and sees and understands a scanned invoice because of the document processing service powered by AI/ML.
The work that was done by a group of employees 8 hours a day can now be done by bots that work round the clock, process invoices, and help the company serve its customers during the pandemic and after.
Low-touch Sales Order Processing with RPA
An Indian conglomerate which makes packaged snacks had major challenges with their sales order process.
The process was manual and involved more than 20 clearing and forwarding agents spending 15% of their time on manual sales order processing. It was slow, with each order taking more than five minutes to complete and it was error prone which resulted in additional efforts for reconciliation and correction.
They solved this problem by implementing RPA. The process flow is as shown in Figure 3.
1. Orders are entered into a web portal
2. Then these files are moved to a shared folder using a scheduled job
3. The RPA bot then takes over
a. It extracts data from excel
b. Creates sales order in the ERP system
c. Sends confirmation e-mail if order is created successfully
d. Sends exception e-mail if sales order could not be created
e. And finally generates the documents for delivery order
The values derived from this automation are as follows:
i. Fast – RPA bot takes less than 30 seconds to process each order, down from five minutes. 10x faster.
ii. Accurate – Eliminated reconciliation and data-entry errors
iii. Efficient – Paperwork for delivery orders generated in tandem with sales orders, and
iv. Time saving – Clearing and forwarding agents can now focus on higher-value tasks. More than 6000 person hours were reassigned.
Automation of Accrual Upload
An Indian Oil and Gas company had challenges in managing the upload of accruals.
Every month end, the finance operations team, comprising 6 users, would carry out the process of accrual. This involved validating excels with accrual data coming from several end users then preparing data in a specific format and finally posting into an ERP system. This manual execution was time consuming and error prone.
The solution was to use RPA to automate this process.
Multiple bots were designed to do the following –
1. Read excels from a folder
2. Refine data based on business rules
3. Validate based on system data
4. Post into ERP system, and
5. Notify a user who then corrects the data if there are errors.
Month end closings are now relaxed, work that required 24 person days of effort is now done in less than an hour and employees can focus on high value tasks rather than spending most of their working hours in email exchanges and manual checks.
Conclusion
RPA generates business value by improving operations to mobilise resources for high-value tasks at lower costs. It increases service quality to reduce cycle times for revenue generating transactions, it increases compliance and analysis capabilities through well-documented audit trails, reduces human errors and helps organisations gain speed and efficiency. RPA with related and complementary technologies such as AI/ML, Process Mining, etc., creates many opportunities for businesses to optimise how they operate and really become intelligent enterprises.
Satyadeep Dey has 23+ years of experience in product development with expertise in Customer Relationship Management, Robotic Process Automation, Machine Learning/Artificial Intelligence and highly available cloud-based services and platforms. He currently leads an engineering team in India which develops SAP Process Automation and also contributes to customer and partner engagements.
https://www.linkedin.com/in/satyadeep-dey-140b2112