If you are tired of performing tedious and repetitive tasks and witnessing too many errors upsetting your clients, your business may need the use of Robotics Process Automation empowered by software robots and the capabilities artificial intelligence.
In this article, my aim is to briefly introduce what Robotics Process Automation (RPA) is, it’s business value, the major use cases, how it works, how we can enhance it, and my experience on architecting solutions to meet the customer requirements.
What is RPA?
RPA is a rapidly growing technology offering which uses software robots and the capabilities of cognitive computing and artificial intelligence. RPA can be ideal for dealing with the legacy systems, modernising processes, and even can contribute to overall digital transformation goals for many business and government organisations striving to cut cost, increase profitability, and delight consumers.
What problems do RPA offerings solve?
At the most basic level (level 1), RPA is used to address repetitive and cumbersome work related problems; for example, tedious clerical tasks taking too long to complete and prone to human errors.
At the second level, it can solve problems related to departmental activities especially for heavily administrative tasks being performed across multiple applications in the business departments.
At the third level, it can solve problems at the enterprise level. Business organisations at the enterprise level can have multiple fragmented or integrated systems which require data exchange. Thanks to the robotic speed and computer accuracy, RPA can solve data exchange problems across heterogeneous systems. For example, customer onboarding process using multiple enterprise applications is one of the problems that RPA can solve efficiently and effectively.
At the fourth level, RPA can also solve problems at an expert level. However, not all RPA products or offerings can be at this level because the expert level requires AI and Cognitive systems integration to RPA core capabilities.
Therefore, when I was analysing customer requirements for RPA solutions, I noticed that only a few products on the market were able to address expert-level problems at the enterprise level. I cover and explain this point further in the article.
What are the common use cases?
There are many use cases for RPA. I encountered the following use cases in my solutions. These samples were generated during the use case collection workshops.
- Security need to be hardened
- Data transfer needed between systems
- Compliance must be expedited
- Products need to be delivered faster
- Matching data between systems
- Integration with legacy systems
- Migration of data between systems with no integration
- Extract data from documents (e.g. email, faxes, paper forms)
- Convert unstructured text to structured data
- No back-end integration available
- Integration is only possible via the UI
- Automate high volume of checks
- Filter out simple cases
- Refer to users for exceptions
- Update fixes and patches faster
- New systems with missing functionality
- Impossible or too expensive to customise
- Workarounds needed
What is the business value of RPA?
One of the most common business value of RPA is productivity, efficiency, and customer satisfaction. While performing some proof of concepts or proof of technology initiatives, we noticed that the software robots could be up to five times faster than an experienced human agent in completing the same task. Not surprisingly, these software robots can work 24/7 basis but human agents cannot.
Accuracy, Compliance, and Security are the major concerns in many business and government organisations. This is a critical value proposition that RPA provides and exceeds in delivery. Our experience demonstrated that there was a substantial reduction in human error rates. We also experienced that RPA enhanced compliance and security capabilities by addressing concerns of many C Level executives in business and government organisations.
As we know business and government organisations have tremendous focus to cut cost, increase profitability, and generate new business. RPA adds a compelling value proposition for cost-cutting, increasing profitability and generating new business. At the highest level, RPA can get rid of downstream process remediation, troubleshooting, and can reposition organisational labour to higher-value tasks.
Last but not least, close to my heart as an enterprise architect, RPA can address major concerns around scalability, agility, and flexibility in organisations leveraging multi-cloud services and platforms. In business terms, this means that the software robots can easily be replicated and rapidly scaled, across the enterprise and beyond, to meet peak or atypical workloads on demand.
How does RPA work at the fundamental level?
The internal ingredients of RPA systems are complex and sophisticated however from a functional perspective it can be easier to describe the operational model by abstracting the technical details. RPA is based on multiple software robots, connectors, and the control plane.
The software robot can read a computer screen. In addition, as designers or operators, we can use pre-built connectors to capture data, e.g. user screen. We can also edit the tasks that software robots produce.
Based on these data collection models, the RPA system can create triggers and timer schedules on the Control Room engine. Then the Control Room engine assigns tasks from the queues to the software robots (a.k.a. Bots) in the worker pool. This is an iterative and repetitive process for desired outcomes.
How else can we enhance RPA capabilities?
If we have a compelling business case and want to use AI capabilities to enhance our services, we need to extend our current RPA capabilities using integrated offerings or customise our solutions at the architectural, design, and specifications levels.
For example, one of my clients wanted to leverage voluminous of streaming data from the Big Data analytics platform operating with a very complex set of open source and proprietary data tools, processes, and technology stacks. The data sources were coming from internal and external sources with high velocity.
For this specific and bespoke solution construct, the only viable integrated RPA service offering with cognitive capability was the IBM Watson suit of products with cognitive analytics capabilities embedded with IBM Cloud Pak for Data to meet this specific requirement. In this bespoke solution, the RPA tool which the client selected was Automation Anywhere. Watson cognitive suit of tools empowered by IBM Cloud Pak for Data met the requirements and the business goals of the organisation.
There are many service providers offering RPA products and services with varying capabilities. I am not in a position to recommend or criticise any products as a vendor-agnostic enterprise architect. However, as solution or enterprise architects we must develop a set of criteria to meet our business requirements and and strategic goals.
We cannot purchase these products and services based on their reputation or market shares. As architects, we must ensure that the product and associated services meet our solution requirements, use cases, financial position, and our organisation’s strategic business direction.
I’d be interested to learn from your experience if you have already architected, designed, operated, or used RPA systems.
This article reflects the personal views and experience of Dr Yildiz.