RPA series: ABBYY - RPA shortcomings in automating content-centric processes

Galusha writes as follows…

Finding a simple technology solution for a complex problem is not always easy. For example, every industry and business department still rely heavily on documents in digital or printed format coming from different communication channels, such as email, fax, mobile, and desktop scanners.

Over the past few years, companies have been using robotic process automation (RPA) systems to streamline these document-related processes, i.e. the input of data into systems. In practice, the RPA deployment has revealed significant shortcomings related to unstructured content and inefficient processes. However, when combined with complementary technologies to turn unstructured data into structured and gain insight into processes, this is where we have seen RPA and its new class of digital workers show great promise.

Limitations with unstructured content

RPA has been successful at automating structured, repetitive tasks associated with processing data across many systems. However, robots can quickly hit a roadblock any time the process involves content that is not structured.

That is very concerning since the overwhelming majority of enterprise content is unstructured. This includes documents, contracts, handwritten information, email communications, and images. Think about some of the common business functions involving back-office financial processes or onboarding a new customer.

These processes always have unstructured content associated with them, which puts an enormous strain on operations and employees to perform tasks like routing documents to the right business groups, verifying the accuracy of data, trying to make a business decision based on information scattered across many different documents, etc.

In order for the new digital workforce to become more efficient, RPA tools are leveraging complementary technology that can learn to recognise, read, and understand all documents regardless of their format – digital and image, structured and unstructured.

Information validation

By leveraging the right visual perception, understanding, and insight skills, a robot can fully automate a process involving documents using technologies like intelligent optical character recognition (OCR), natural language processing (NLP), and machine learning (ML), to digitise content, classify documents, extract data and validate information with little human involvement.

Vision, understanding and insight skills are central to content intelligence (IQ), where AI is applied to:

• Easily automate and analyse content-centric processes involving images, documents, texts, and communications.
• Analyse and learn from content to make more informed decisions.
• Incorporate machine learning to perpetually improve and streamline business processes.
• Measure, sustain and adapt digitised content processes over time.

Smarter RPA bots

To grow and expand the use of RPA within an enterprise, robots must become smarter to be able to interpret and understand unstructured content and turn it into actionable structured information. Think of RPA as the starting point for intelligent automation, then with the addition of content IQ skills, digital robots can have varying degrees of intelligence.

For reference, content IQ is defined as a class of enabling technologies that help digital workforces understand and create meaning from enterprise content. Content IQ provides the ability to automatically extract all relevant information from documents and break down the processing of content into easy to use and consume technology. It can be leveraged directly within an automation solution, targeting activities and skills required to solve specific business problems.

Connecting content IQ skills is easy. They can be integrated into RPA platforms like UiPath and Blue Prism using pre-built connectors. Content IQ skills are like services, they are very extensible, so that enterprises, partners, and software vendors could integrate them with any business process automation system.

When combined with RPA, content IQ bridges the cognitive gap digital software robots have when it comes to unstructured content.

Insight into RPA process workflows

Another common challenge associated with RPA is that it requires process transparency. If you don’t take a process-first approach, it can be like building a LEGO set without directions. Knowing where to start with RPA as well as the benefits and risks to expect before, during, and after implementation are all elements that need consideration.

Realistic expectations should be set.

Adding process intelligence (IQ) to automation efforts gives the foresight needed to improve the success of an RPA project. From selecting the right process to automate with cost analyses, quantifying the effects of processes downstream, and monitoring digital worker performance post-implementation, process IQ technology in a way acts as a crystal ball on how RPA will behave.

RPA ignited the fuel to accelerate digital transformation. Now, with the addition of content IQ and process IQ, organisations can achieve significantly better results with their initiatives.

Galusha: RPA requires process transparency.