Adding Cognitive Fuel to Supercharge Your RPA Deployment


No matter the industry, the chances that robotic process automation (RPA) can accelerate turnaround and save costs in the back office is well documented. Some companies are seeing a return on investment as high as 300 percent. It ends up that deploying a relatively inexpensive software robot to take on tasks previously performed by humans has loads of upsides. Eliminating human error alone can save a company millions of dollars a year.

Now, many companies that have pioneered RPA are looking for their next move. They want to know if – and how – they can extend the benefits they are reaping from their RPA deployments. Can cognitive technologies supercharge their automation efforts? And will the business case still stack up if they use multiple technologies?

The simple answer to both of these questions is yes. But the artificial intelligence (AI) and cognitive technology landscape can seem like a rainforest of discovery, teeming with different technological species. It’s an easy place to get lost – especially if you don’t know which specific business problem you are trying to solve. Choosing the wrong cognitive solutions for your environment can be like chiseling granite with a knife.

An important and difficult first step is to filter out the noise from the market and understand the problems these technologies can actually solve. Many of the solutions being touted today as AI do not actually learn or solve problems, as the name might suggest. However, they are capable of using methods based on probability and economics that mimic “intelligent” functions. For a buyer, the semantics matter.

The second step is to determine how cognitive and AI technologies can build on the RPA capabilities in the market today. Take a process that involves gathering data from scanned images, like invoicing, for example. Until recently, capturing accurate data from scanned images has required human intervention. With a cognitive optical character recognition (OCR) engine that digitizes printed texts so they can be electronically edited, searched, stored and used in automated processes such as RPA, those scanned images are suddenly a rich source of digital data. And the OCR engine allows RPA to pick up its tasks in the process without a person setting off the sequence. Similar results are possible with natural language processing technologies, which can process unstructured text in a written email, identify the intent of the text and pass the rest of the process on to an RPA bot to automate the succeeding tasks.

Because cognitive technologies create a point in the process from which RPA can take over, more tasks in the process can be automated to be more efficient, and more data can be used to generate business value. Put simply, without cognitive technology, many processes don’t warrant automation with RPA alone. With cognitive technology, the “automatable” piece of the pie gets bigger.

The advent of RPA has brought many benefits to the enterprise – automating rules-based and repeatable processes to make them faster and less prone to error are the most obvious. But has it – in the meantime – changed the way newer cognitive technologies have been designed? Is “business led” automation starting to creep into areas previously designated only for highly trained and costly IT folk? IBM’s Watson Virtual Agent is a natural language interface chatbot that can be trained with customer service subject-matter experts, meaning no costly IT programmers are needed in building the customer journey.

Indeed, the lines between IT and business operations are increasingly blurred, with more and more technology finding its way into organizations through operations. This is great news for businesses of all sizes. Business-led automation tends to be cheaper to implement and has many success stories, especially in RPA.

So is investing in cognitive solutions worth it? Early indications show that advances in technology mean that upwards of 35 percent of the processes that have been considered “un-automatable” are now candidates for automation. Even if cognitive technologies cost more than RPA, they can solve bigger problems and – if implemented in the right place in the right way – yield higher savings.

It’s an exciting time to explore cognitive technologies – but the path can be uncertain. Join me at the ISG Automation Summit, September 19 in Amsterdam – or my colleagues at the ISG Automation Summit, July 10-11 in New York City – to explore the cognitive opportunity.

Or contact me directly to discuss how ISG can help you along your automation journey. 

About the author

Wayne is an automation pioneer, initially starting out as an early adopter of RPA in 2010, creating one of the first Enterprise scale RPA operations. His early setbacks at Telefonica UK, led to many of the best practices now instilled across RPA centers of excellence around the globe. Customer centric at heart, Wayne also specializes in Customer Service Transformation, and has been helping brands in becoming more Digitally focused for their customers. Wayne is an expert in Online Chat, Social Media and Online Communities, meaning he is perfectly placed to help take advantage of Chat Bots & Virtual Assistants. More recently Wayne has concentrate on Cognitive & AI automation, where he leads the European AI Automation practice, helping brands take advantage of this new wave of automation capability.