Skip to main content

Deciphering data – how to meet this RPA challenge

(Image credit: Image Credit: MNBB Studio / Shutterstock)

Organisations increasingly need to unlock data from Documents and systems of record, to increase ROI from RPA. This challenge of data acquisition and transformation is now being solved by connecting robotic process automation (RPA) with other embedded and integrated cognitive technologies.

Without the ability to capture semi-structured and unstructured input data, from documents and other sources, the true potential value of RPA remains unfulfilled. It would be great if we operated in a world where all data is in digital form - at source - and is at least semi-structured, but unfortunately we don’t. The reality is that many processes are still “paper based” in some form. Even when digitised, documents are often in highly unpredictable and variable formats – differing widely in quality, structure and complexity - bank loan applications, invoices, and expense reports, for example. They may contain both printed or handwritten text and are often packed with tables too.

We can also add to the document capture challenge, all the huge volumes of free form text from fields in systems of record, chatbot interactions, emails and a myriad of other interfaces. Ultimately, there’s a mess of data out there that often hampers an organisation’s ability to transform its processes with traditional, rule-based RPA.

To see how prevalent this problem is, we’ve analysed the statistics from our Digital Exchange (DX) – which is a marketplace for accessing and downloading pre-built artificial intelligence cognitive and disruptive technologies. This insightful data revealed that over 60 per cent of downloads from the DX are ‘visual perception’ skills that relate to acquiring or understanding data. This confirms that most organisations are now looking at how they can expand beyond simple, structured and rule-based processes.

Current solutions

Transformational, technologies like connected-RPA, provide the ability to address these challenges, by using automated, digital workers to emulate repeatable tasks that would otherwise be performed by human, software users. What makes connected-RPA software so compelling is its ability to connect and navigate applications and processes in a highly flexible way - without needing to rely on the IT department to write code and build integrations and applications.

Business users simply create automated processes in a “visio-like” designer, which are then used by the digital worker to automate a task – using the same applications. This technology is evolving – by combining more embedded intelligent automation capabilities and a tighter integration of people within the workflow - to enable collaborative innovation across the enterprise. This is achieved by giving business users the ability to easily access and exploit leading-edge cloud, AI, cognitive, and other capabilities, through drag and drop, intelligent automation skills.

Addressing a common use case

A large European multinational is addressing a number of key finance operations process challenges – specifically across its ‘source to pay’ activities. With huge volumes of invoices to process each year, it made sense to help staff to better manage invoice processing – which is very difficult when they’re in multiple languages and locations. This meant simplifying the complexity of these tasks, removing systems bottlenecks and errors - especially at peak times – while reducing the overall cost of service delivery too.

Connected-RPA, provides the ideal solution to address these challenges, by using automated, digital workers to emulate repeatable tasks that would otherwise be performed by human, software users. This eliminates low-value, manual interventions at the interfaces between systems and process steps.

This activity involves digital workers automating the extraction and processing of data from digital documents; classifying different types of documents, providing quality checks, detecting errors and passing exceptions to humans and managing the data flow and accuracy. Another ‘month closing’ finance process is being managed by a single digital worker, which has already reduced completion time - from hours - to minutes.

Moving RPA forward

Unlocking data is the key to unlocking greater and more complex automation and improved customer experience and doing this holistically is a highly complex problem and requires an evolution of connected-RPA – with a combination of embedded and integrated technologies.

To support this next evolutionary wave of automation, certain capabilities need to be considered as “table stakes” within an Enterprise RPA platform - the ability to ingest, process and understand data is becoming one of those capabilities. RPA is evolving to support these additional capabilities – connected-RPA makes these more complex problems accessible, by incorporating traditional technologies – such as OCR – with supervised and unsupervised Machine Learning, to enable a business-led approach to training digital workers to understand new documents and data structures. Doing this requires that we bring humans more into the process, as training the digital workers on these new problems requires supervision and / or up front training.

Connected-RPA is increasingly seen as providing the connecting ‘glue’ to integrate other cognitive technologies and APIs. We only need to look at the new offerings from the major cloud vendors - such as Microsoft with its Form Recogniser and AWS Textract - to see the impact that AI is having on this Data extraction space and the potential for commoditisation of some basic form extraction capabilities.  However, this is a complex problem. There will continue to be difficult to solve problems that will require a more specialised approach – complex or domain specific documents, poor quality images, or handwriting recognition – for example.

By combining a built in, standard Document Processing workflows – optimised for a business user - with easily integrated complementary technologies and commodity AI, such as Natural Language Processing and Translation, businesses can create end-to-end transformational process automations that deliver increased value to the business and deliver improved customer experiences.

Final thoughts

We are at an early stage in this journey and connected-RPA will continue to evolve to solve more complex problems. Through an embedded, machine learning enabled, business-led user experience and connected-RPA strategy – organisations will soon be able to create a recipe for deciphering the data and language of their business. Ultimately, this will usher in a new wave of Intelligent Automation and will allow them to free up those ‘digital entrepreneurs’ that exist in those organisations to work on higher value tasks.

Colin Redbond, head of technology strategy and architecture, Blue Prism (opens in new tab)

Colin Redbond is head of technology strategy and architecture at Blue Prism, and is a recognised thought leader in RPA – with over 20 years experience in the IT industry.