Robotic Process Automation gained traction couple of years ago with a lot of hype about eliminating keystrokes in offices. And it is true that we all spend far too much time sitting at a keyboard. The catalysts were:
- expensive financial analysts sitting screen scraping from corporate filings to paste the data into their Spreadsheets and
- data entry outsourcing companies facing rising labor costs.
Much time is spent using keyboards to move data from one format to another, one spreadsheet to another and from one computing platform to another. HSA calculated that in manufacturing alone companies are spending $1bn a year on simple desktop key entry in the US – most of which could be automated with RPA.
Capture has always been about eliminating key entry from paper, fax and emailed forms – We’ve been dealing with it for many years. HSA data shows that Capture represents a market that was worth about $4bn in 2018. RPA represents a small sub-segment of this today – but it is rapidly growing.
As users began to automate key entry processes at a single user level by building ‘bots’, they rapidly found that they needed to start to incorporate paper and “paper-like” forms such as PDF’s. Although many times commonly assumed, all input to a process is notnicely structured electronic data. For many automation projects users found that scanning and OCR and other pattern recognition and classification technologies were required.
Associated with extracting and validating the data from these forms types is business process workflow. This resembles a lot of the ECM and old style back office Batch Capture processes that HSA has been dealing with for years. We call that Capture 1 systems – it still represents a large market. What is new is using AI and machine intelligence to classify the forms types and identify the data that needs to be extracted with associated rules. Originally we did this with templates, but those are long gone. Today most traditional capture vendors do this via OCR and proximity searching e.g. if a field type was labeled Invoice Number, then probably the Invoice Number that has to be extracted was located in the labeled box. New AI based Intelligent Recognition builds up an understanding of the forms types with layouts, extraction areas and recognition rules which can obsolete the old methods
HSA calls the series of services associated with automated understanding, classification, and extraction of corrected valid data, Capture 2.0. Maybe it is more properly termed Intelligent Capture 2.0, since leverages machine intelligence. Capture 2.0, which we calculate as about a $30bn services market, consists of a series of micro services to perform such activities as intelligent classification with deep learning, extraction, and validation or purification of the data as well as the ability to perform high speed key entry repair or correction – even double key entry where 100% accuracy is required. It includes such areas as Natural Language processing (NLP) and language translation as well as semantic understanding and sentiment analysis. These smart technologies can be applied together or separately, these services “building blocks” provide the foundation to eliminate time-consuming and expensive data entry to feed faster processes.
Smart is a difficult word – it may be interpreted many ways. Webster’s first definition is as ‘causing a sharp stinging’. Its second is“marked by often sharp forceful activity or vigorous strength” and third is“Brisk, Spirited” – only the fourth definition is “mentally alert or knowledgeable”. Intelligent RPA becomes knowledgeable about its inputs, what is needed for which processes and what to do with the extracted information.
The fact is that we are ‘capturing’ the key data from multi-channel inputs derived from many sources in multiple different ways. What is changing is the level of data understanding and integration to the back end processes leading to “smarter” or maybe “Intelligent” Capture 2.0 automation services.
