The above image from Workfusion’s Combining RPA + AI webinar, nicely summarizes both RPA and cognitive automation. While RPA is the doer, cognitive is the decision engine.
Cognitive automation or also called intelligent or smart automation is the hottest field in automation. Talk to any RPA company CEO and they will start talking about cognitive automation, at least Max from WorkFusion was most excited about cognitive automation in our podcast.
While automation is old as the industrial revolution, digitization greatly increased activities that could be automated. However, as you can read more on our guide on RPA tools, initial tools for automation, which includes scripts, macros and robotic process automation (RPA) bots, focus on automating simple, repetitive processes. This makes sense because most core corporate processes are quite repetitive but not repetitive enough to completely take human out of the loop with simple programming. However, as those processes are automated with the help of more programming and better RPA tools, processes that require higher level cognitive functions are next in the line for automation.
Key capabilities for cognitive automation
Natural language processing (NLP): Even basic language understanding makes it much easier to automate most customer service processes or processes involving contracts.
Optical Character Recognition (OCR): Despite increased digitization, a mind-bogging amount of paper is still used, especially in heavily regulated industries like healthcare or banking. Processing these paper are required to automate any process end-to-end.
Machine learning: Processes require decisions. If those decisions can not be formulated as a set of rules, machine learning solutions are required to replace human judgement with machine judgement and automate processes.
Let’s look at what bots will do with these capabilities.
Discovering mismatches between contracts and invoices: Deloitte explains how their team used bots with natural language processing capabilities to solve this issue.
Offering end-to-end customer service with chatbots: While chatbots are gaining popularity, their impact is limited by how deeply integrated they are into your company’s systems. For example, if they are not integrated into the legacy billing system, a customer will not be able to change her billing period through the chatbot. Cognitive automation allows building chatbots that can make changes in other systems with ease.
Banking – Fulfilling KYC requirements: Leverage public records, handwritten customer input and scanned documents to perform required KYC checks.
Banking – Processing trade finance transactions: Banks finance international trade. Processing these transactions require paperwork processing and completing regulatory checks including sanctions checks and proper buyer and seller apportioning.
Insurance – Servicing policies: Data mining and NLP techniques are used to extract policy data and impacts of policy changes to make automated decisions regarding policy changes.
Insurance – Claims processing: Make automated decisions about claims based on policy and claim data and notify payment systems.
Though cognitive automation is a relatively recent phenomenon, most solutions are offered by Robotic Process Automation (RPA) companies. Check out our RPA guide, most comprehensive public repository of RPA vendors or our guide on RPA vendor comparison for more info.