Conversational Data Intelligence and Optical Character Recognition (OCR) are complementary technologies. When combined, they provide business leaders with all the tools they need to deliver complete email analysis and end-to-end automation.
Email automation is one of the most popular use cases for Conversational Data Intelligence. A platform like Re:infer can help enterprises extract and understand the content of their mass communications, providing the data needed to automate common and time-consuming email-based tasks. This includes everything from routing and manual triage to Communications Mining and opportunity discovery.
However, email attachments introduce a new layer of complexity to the process. Conversational Data Intelligence is optimised to understand people expressing themselves in natural language. But comparing this to the information contained in documents or images is like comparing apples with oranges - interpreting data in attachments requires different machine learning techniques. Conversational Data Intelligence can make predictions based on attachment names, but it isn’t designed to read the contents of attachments.
Fortunately, a simple and fast integration with Optical Character Recognition (OCR) technology allows Conversational Data Intelligence to close the loop and enables complete email automation.
OCR and IDP
OCR is a class of enterprise technology that converts text in image files into structured, machine-readable data. It can be used to extract data from scanned physical documents as well as text in images, for example - when an iPhone reads text in photos. OCR is also a critical component in Intelligent Document Processing (IDP) solutions. These combine OCR with AI technologies - like natural language processing - to convert unstructured and semi-structured information into usable data. This includes the vast majority of file formats, including PDFs and electronic text documents.
Between OCR and IDP, businesses have all the tools they need to rapidly extract insight from almost any email attachment. Yet these have limited value without first understanding the content of the email itself. This is the role played by Conversational Data Intelligence, and it’s why these complementary technologies work best together.
Integrating OCR and Conversational Data Intelligence
A complete email automation solution needs three core components: a trained Conversational Data Intelligence model to understand the email, IDP technology to read attachments, and an automation tool to handle business logic and tie everything together. This could be a serverless engine, a service you deploy yourself, or even a simple RPA.
Once you have these components, you can start configuring your solution. Firstly, the automation tool will need to receive the structured interpretations of emails from the Conversational Data Intelligence model. This should be done in a way that provides your tool with a real-time stream of structured communications data, which can then be automated as your business requirements dictate.
A Conversational Data Intelligence tool will help a business decide if an email is in-scope for automation. But the Re:infer platform is unique in that it also tells you if there are any attachments present in the email and what their names are. Your automation can use these attachment names to decide whether they should be sent for processing by OCR.
Once this is done, all entity predictions from each email in the thread and every data point OCR has extracted should be aggregated. Your automation will then make a judgment on whether it has everything it needs to process the email. If it does, it will progress to end-to-end automation. If it does not, it could go back to the user to ask for whatever is missing, or send the email for manual validation.
Closing the loop
Integrating an OCR tool with a Conversational Data Intelligence platform like Re:infer expands the scope of automation in your enterprise. By combining the former’s understanding of email with OCR’s ability to read attachments, you can automatically extract every nugget of information an email contains. With understanding and insight then comes action, and the ability to create powerful, automated email workflows that drive efficiency and competitive advantage.