The Rise Of Cognitive Robotic Process Automation
Cognitive robotic process automation (RPA) is a fast-evolving field of computing and is an emerging form of business process automation (BPA) technology. It involves the automation of many internal and external customer journeys through software “bots.”
Where RPA Started
RPA started roughly 20 years ago as a rudimentary screen-scraping tool, technology that is used to eliminate repetitive data entry or form-filling that human operators used to do the bulk of. For example, the software could copy data from one source to another on a computer screen. Imagine a finance clerk handling invoice processes by filling in specific fields on the screen. Early RPA was able to take this function off the clerk’s plate by automating that invoice processing.
The insurance sector soon discovered how this technology could be used for processing insurance premiums. Typically, when brokers sell an insurance policy, they send notices using a variety of inputs, such as email, fax, spreadsheets and other means, to an intake organization. Intake teams historically managed this multi-step sales process manually, including organizing the data, checking for completeness and accuracy, working with brokers to correct errors, extracting other necessary data from online sources and then completing the sale.
Insurance intake teams and operations teams have, in the last few years, used RPA software to run the structured parts of the intake and claims process. Specifically, these teams would organize incoming data and then feed that data to back-end software bots. The bots would then collate this information into systems of records to complete the workflow. All of this was happening around 2016.
Cognitive RPA Brings Intelligence Into The Equation
Thanks to recent advancements in artificial intelligence (AI) and machine learning (ML), process automation has evolved from being a mere screen-scraping technology with bots handling repetitive processes to a technology that is more cognitive in nature, enabling software bots to make intelligent decisions that assist human workers.
Cognitive RPA (CRPA) involves technologies such as natural language processing, machine learning and deep learning that take information already available in the enterprise to create models that lead to autonomous, cognitive-based decisions. This entails understanding large bodies of textual information, extracting relevant structured information from unstructured data sources and conducting automated two-way conversations with stakeholders. A good application for CRPA is taking accepted and rejected insurance applications and feeding them into a system that can learn how those decisions were made based on information in the applications. CRPA software is then able to automate the acceptance or rejection of subsequent applications, leading to considerable cost savings for the company.
CRPA Systems In Action
The insurance sector is just one vertical segment that’s taking advantage of CRPA technology to expedite the claims process. One company we’re working with told us their agents were making more than 650,000 outbound calls per year in their attempts to close short-term disability claims. These agents were making, on average, six call attempts to reach a claimant to get the required information needed to close the claim. This was a manual process that took three weeks and about $17 per call.
After implementing CRPA into their system, the company built conversational and process paths into their claims systems that automated connecting with claimants using two-way text messages. They were able to build these new claims workflows in a few days. In the end, the company reduced the claims processing time from three weeks to one hour, saving the company roughly $11.5 million.
The perks of CRPA are boundless and applicable to all industries. Another vertical segment taking advantage of cognitive automation is the manufacturing industry. Chart Industries, a manufacturing firm within the energy sector, utilizes CRPA to enable their accounting division to be more efficient and cost-effective — a use case which any business in any industry can capitalize on. Chart allocated multiple different back offices to handle accounts payable, accounts receivable and other tasks, resulting in unaligned processes and procedures.
To maximize efficiency, Chart Industries deployed a process automation vendor, Celonis. Using machine learning to identify patterns and irregularities, Celonis’s technology identifies business accounting processes and determines and performs the corresponding processes. Chart was able to save an annual amount of $240,000 from late payments alone.
To learn more about the return on investment (ROI) of CRPA, I recommend reading “Understanding RPA ROI” by the Institute for Robotic Process Automation & Artificial Intelligence (IRPAAI).
Choosing A CRPA Platform
My next article will examine how to choose a CRPA platform in greater detail, analyzing features and functionalities organizations should evaluate to deliver straight-through processing. But for those who want a taste of the process, it's important to identify and assess the following key features in a potential solution:
- Data Security: When outsourcing sensitive customer data to business to business (B2B) companies, it is imperative to safeguard your customers’ information. Ask any potential vendors if they are SOC 2 compliant. SOC 2 compliance is a voluntary and stringent auditing procedure that probes the security practices of companies storing customer information in the cloud. It guarantees that only the most trustworthy and airtight data security practices are being used.
- Omni-channel Capabilities: Companies and their technologies vary widely — some companies can automate processes over APIs, documents, SMS, voice, email and/or the web. Assess what your company needs and values. Email triaging? Chatbots? Voice recognition? Intelligent document processing?
- Low-Code And No-Code Integrations: Business process transformations that are often dependent on IT teams and having solutions that can deliver outcomes with minimal implementation efforts lead to considerable ROI quickly.
- Ability To Digest Unstructured Data: A large majority of all data is unstructured and does not fit into a predefined mold. Ingesting unstructured data increases a company’s CRPA potential exponentially, as their models can handle any and all data that is thrown their way.
When choosing a CRPA platform, it is important to take all these factors into account. Due diligence at the beginning of your implementation will make sure your automation initiatives result in quick efficiencies and ROI.