Fast is fine but accuracy is everything.” Wyatt Earp

Wyatt Earp an old American lawman was probably referring to OCR Optical character recognition when he shared this quote with the world or was, he referring to other things we don’t know. But guess this summarizes the essence and role of OCR.

Nearly about 80% of the F&A, Procurement, supply chain organizations, governmental organizations begin their automation journey from existing paper-based processes. This probably is the first step towards digital transformation. Quality is a critical issue here or rather the accuracy of the word or character matters. It makes a difference sometimes maybe even running into reams of documents.

Varied scenarios

Data can be handwritten like cheks and customer on-boarding forms, unstructured like contracts or semi-structured like invoices and receipts. This cannot be processed straight through to get the output. There must be accurate context setting, reading and interpretation level capabilities associated to get it to better. Organizations that use IDP - leverage OCR to read this data, classify and then verify it before ensuring it is in the needed and understood format. The very process of IDP document processing involves observing human action, self-learning and then moving into an automated mode to handle volumes of data. In addition, the cognitive engine and machine learning algorithms help to recognize new formats on the fly.

Automation from Jiffy

JiffyRPA powered by Option3, an automation and data analytics product company are focused on providing niche automation solutions for companies, big or small, across the globe. The company empowers businesses with AI-powered Intelligent Automation solutions to solve complex business problems across industries and functions. Its choice of technology, a great combination of the Intelligent Automation along with Intelligent Document Processing with OCR and Smart Analytics forms the trifecta and couple it along with People and Processes, Jiffy provides the single stack automation hub with Human-in-the-Loop.

OCR does require certain criteria to make it work well. For example, the image quality, resolution that must be set moderately, quality of the scanner, handwritten content, background and colour images, scanning text on scanned images, etc are all to be checked upon.

Jiffy comes in with its ability to integrate seamlessly with any third-party OCR in addition to inbuilt OCR support that helps to perform to its maximum. It can also add on more OCRs along with multi language support, ability to recognize new file formats and a data interface that helps in optimal performance.

Probably the earliest forms of OCR were creative reading devices for the blind and needless to say, the largest use of OCR is in reading invoices, other than - Data management, Format standardization, KYC verification, Account servicing, Query handling, Responding to customer queries both regarding their account and other banking products, Contract management, Claims processing, Competition analysis, Payments processing, Fraud detection, etc. Like for example in the case of contracts, pre and post contract documents are saved in different folders for comparison and the match results run back in JDI and the final approved files in a matched folder.

Mitigating Risks by Mitigating Errors

Accuracy is achievable because it can be measured and anything that can be measured must be close to its true value. As there is a true value, Technology hinges on the fact that it helps build processes and tools to achieve that number.

The human variable is the riskiest one in data entry process. With OCR software that variable is all but removed from the equation. The result is consistent and pre-validated data for the organization and a much lighter workload for its staff, who can be allocated for higher-value tasks.

Importantly, the OCR engine does not tire or lose its concentration after hours of repetitive, detailed work. It remains as accurate after the millionth form as it was after the first (in fact, more accurate, thanks to repeated “learning” over time). And that gives accuracy to the final output and learnability ensures the system remains intelligent!