Traditional software development builds code to capture best practices and embodies them into the back office of enterprise applications. Robotic process automation captures the keystrokes of the workforce on the front end of applications, as they work and deliver, and in that sense automates them into software robots or bots.
This process of automation reduces the barrier to automate by less rigorously involving software coding and application programming interfaces and has delivered its share of rapid success over time.
As building a software robot using RPA tools is relatively quick, adoption of RPA and delivering the first set of robots for any organization has been relatively easy. RPA as a technology is robust enough to handle complex process automations and scaleup quickly, but it is when we need to solve complex problems that requires mimicking human brain is where we hit a roadblock with RPA.
Arrival of intelligent automation
Till now, RPA has unleashed a digital workforce of software robots that has worked and delivered on the periphery of much wider enterprise initiatives such as digital transformation. But in order to move forward, RPA will now need much deeper and better integration with subjects such as machine learning and computer vision.
This is leading to an evolved form of automation, with RPA at its core, termed as intelligent automation.
Let’s take a very popular example of “Mathematics of Cucumbers”. Cucumbers in Japan comes in a dizzying variety of sizes, shapes, color and degree of prickliness and based on these visual features, they must be separated into different classes that demand different market prices. A machine learning algorithm was written that could classify these cucumbers based on a photograph. A simple problem requiring intense human labor was solved easily using intelligent automation.
The rapid post pandemic acceleration of digital transformation initiatives has also highlighted the need to bring Intelligent Automation into the folds of the enterprise-wide transformation, versus its previous peripheral and sideline approach. The post pandemic pressure on organizational performance and efficiency has meant that automation is now increasingly moving into the core of post pandemic digital transformation.
Learning points
Some of the learning points from the usage of Intelligent Automation: –
Process
According to Forester in its report Ten Golden Rules for RPA Success, May 2020, more than half of all global RPA programs use less than 10 bots. According to Forester, less than 19% of RPA installations have reached an advanced stage of maturity. Some of the setbacks that have stalled programs include fragmented initiatives, multiple vendors and incomplete governance.
Technology
Using RPA, business users can introduce automation across workflows and automations. But often this includes scripting and the more complex the process the more scripting that is required. Bots or automated processes breakdown when scripts fail and these are affected by infrastructure, software, data, and so on.
Culture
Intelligent Automation is an enterprise-wide solution and also need to be managed by change management, use cases, security practices. While Intelligent Automation is an initiative, building an automation culture is important in the longer term. Since automation impacts people, they can either resist or become a strength.
Next level of automation
For any type of automation, it may be necessary to look at the entire process to plan for straight through automation. Process mining can help to streamline and automate the process faster. While digital work assistants can be used for simple processes, for more complex processes it may be required to use task analytics, design thinking, journey visioning. This helps to map user behavior, motivations, dependencies. Once completed, the organization can have a much better view of short- and long-term automation opportunities.
Bots with embedded machine learning can learn to ingest and extract information from multiple data types including structured, unstructured, image, inferred. And from various sources including web forms, PDF files, emails, scanned images using algorithms, and execute tasks involving logic or routing.
Bots with embedded machine learning can manipulate data, perform fetch or fix operations to fill data gaps. They can compare data from multiple sources, to validate downstream action. They can offer data-driven signals to RPA bots, which then take action based on these signals. They can trigger dynamic process changes or support human workflows by anticipating process exceptions. The lists of use cases and adaptations is constantly growing.
Future success
As an organization blends humans, bots and machine learning into processes, the benefits and gains and further opportunities will keep growing, However, in order to be successful, it is also important to build a culture that recognizes and prioritizes automation.
For example, Intelligent Automation could automate that uses machine language to handle exceptions needs humans to train the algorithms, validate results, and manage process exceptions. Automating processes gives an organization to rethink the legacy of its processes and refocus on customers and employees.
Prioritizing automation does not mean that employees and humans are not centerpiece for the organization. There are huge benefits that humans can gain by skillfully blending automation into the digital enterprises and training data and blending machine learning into processes.
Discussion about this post