Robotic Process Automation (RPA) as a transformation lever has seen widespread adoption across most industry verticals over the last few years. Most enterprises have carried out a Proof of Value (PoV) to establish that the technology works, many have automated the low hanging fruit in process areas such as Finance and HR, and some of the mature ones have setup set-up their Automation Center of Excellence to evangelize the technology within the organization and drive best practices.
And yet, achieving scale and infusing automation across all business units continues to be elusive. That could begin to change in 2020 as organizations make the jump from RPA to Intelligent Automation.
The following are some of the key technology levers that will define this shift towards infusing “intelligence” into automation initiatives:
Artificial Intelligence and Machine Learning for cognitive decision making:
Enterprises can use AI/ML algorithms to improve the ROI on existing process automation. The objective of this “AI fabric” is to reduce the number of exception scenarios that require a “human in the loop” to step in. The better the exception rates, the higher will be the adoption of automation.
Take the classic case of Accounts Receivable process. Achieving straight-through processing (STP) is key to the automation of this process. Using deterministic business rules is not enough as often key customer information like invoice or customer reference is missing in payments received. Using multiple ML algorithms for predictive matching based on historical data, and then routing it to an RPA solution for updating an ERP system can significantly improve the STP rates.
Cognitive capture of unstructured data:
Data is the bedrock of business process automation. Most data, however, is in unstructured or semi-structured format – documents (PDFs, images), emails and even voice. Conversion of this unstructured data to structured data for use in automation initiatives requires the use of technologies with the ability to classify, interpret and extract key information just like a human would.
Traditional OCR technologies fall short of the ask and organizations must look at an AI-based approach for their data capture needs.
Typical use cases in the enterprise that can benefit from this cognitive capture are ones that have a reliance on document processing – invoice processing, legal process automation and contracts management, HR processes.
Conversational AI-based customer support:
Enterprises competing with digital native businesses are now putting customer experience as a top priority. Meeting the customers at their channel of choice and adapting to their modes of interaction is key.
Conversational AI leveraging technologies like Natural Language Understanding (NLU) and Natural Language Processing (NLP) combined with sentiment and intent analysis is driving the next-generation customer experience. Conversational interfaces via text-based chatbots or voice can help define new customer journeys ranging from product discovery, customer service, and support and transactions.
Industries such as Banking, Insurance, and Retail have been leading this change, but most other verticals are now adopting conversational interfaces.
As an alternative to traditional call centers and IVRs, conversational interfaces can provide cost efficiencies up to 10X. But perhaps the bigger prize is the ability to use analytics on the interaction data to gain real-time insights into customer behavior, eventually leading to faster conversion and increase revenue.
Process Intelligence tools:
As the focus shifts from task-based automation to automating end-to-end processes, solutions that help with process discovery and creation of a “heat map” to identify where bottlenecks lie will become important. A complete overhaul of the business process is rarely successful. Instead, organizations should follow a more bottom-up approach to identify the high impact process steps and infuse automation using them as “digital hooks”.
An equally important aspect of Process Intelligence is the ability to maintain the context across the automation steps involving a heterogenous environment of technologies and systems (RPA, AI/ML, BPM and legacy ERP, CRM and transaction processing systems”). Automation is not an “either-or decision” between an API/BPM approach and an RPA based approach.
Organizations that adopt solutions to deal with this heterogeneity and use Process Intelligence Solutions to inform their transformation roadmap will be clear winners.
Will 2020 be the breakout year for enterprises looking to remove the friction in their business processes and drive superior customer experiences? The right set of tools are certainly available for this to happen; whether the adoption keeps pace is to be seen.
Originally appeared in AI Time Journal: