By Ravnit Kohli
When you hear the word “automation,” you may picture large, seemingly autonomous machines assembling cars on a factory floor. While this form of robotics has become a surging, multi-billion dollar market, there’s been a definitive shift to robotics helping the average consumer and business. In fact, consumer automation is expected to become a $1.5-billion market by 2019 and may even help automate important business processes.
Robotic Process Automation has become a key priority for bank executives looking to do more with less. When considering the automation opportunities offered through AI, many banks have identified on-boarding and know-your-customer processes as the priority areas. Traditionally a paper-intensive, and employee-heavy operation, the customer acquisition process is often the customer’s first impression and subsequently the prime focus of digital-only banks that seek to disrupt the industry.
This could lay the foundation for automation in much more complex and high-value business actions. Complex processes like identity verification and even fraud prevention could soon become an automated robot’s job.
As automation becomes more common, businesses should consider these four things to take advantage of the opportunity:
– Business requirements – As with any new project, the company should think about whether the problem they have is worth solving to begin with. Will fixing it increase efficiency, save cost, reduce fraud or deliver repeatable value to the business? The answer to these questions must be “yes” to proceed.
– People processes – Robotic process automation’s goal is not to replace humans. People will still be needed to review the auto-generated work, and as the workflow is re-engineered, these people will be freed up to focus on higher-value activities.
– Operations – Once a workflow becomes automated, how could that impact other functions or departments? Businesses should consider how to extract further economies of scale from automated processes across the business.
– Technology – The key consideration to keep in mind on a technical level are:
– Data: It can come in many forms: structured or unstructured. An example of structured data could be numbers in an Excel spreadsheet. Unstructured data might be scanned, written text and numbers that differ across situations, thereby more difficult to understand. RPA is generally capable of processing these different types of data – including screen scraping (that is, copying data from a website), parsing documents and emails, image recognition, accessing a server or website.
– Workflows & Rules: The ideal candidate workflow for RPA is when rules are unambiguous. Once the RPA solution has been trained to capture and interpret workflow processes and business rules, it can then manipulate data, trigger responses and communicate with other systems autonomously.
– Level of Intelligence: The level of intelligence would decide the maturity of the RPA solution especially from an AI point-of-view. RPA solutions vary: from zero intelligence (basic automation) to pattern matching (intermediate level), to sophisticated machine learning algorithms (including neuro-linguistic programming and cognitive based computing to emulate human-like decision making).
RPA tools can be a first step toward more intelligent, automated data solutions. By combining robotics with other AI techniques we get innovations like self-driving cars. But to foster this, businesses need to understand the range of AI technologies and how to translate that technology for the most value. This requires deep technical expertise and understanding of business processes. Strength in one area can only get the business so far.
(Top photo: By Carl Heyerdahl.)
