WHAT IS NEW?
New areas for Document-to-Data-to-Decision automation
Doc Element Identifier – Reduce Training Time and Cost by 80%
Ability to segregate a document into a logical parts is critical to reading and understanding it. This applies to bank statements, salary slips, financial statements, tax returns, invoices and many other such documents. IN-D’s proprietary Doc_Element_Identifier (DEI) is the basis of our best-in-class accuracy and the ability to onboard new documents in a week or less. Among other things, DEI separates tables from free text, organizes tables into rows and columns and identifies their headers. It also helps in ensuring that all our solutions are template independent. We filed our initial patent application last month.
Contract Review and Reconciliation
Verification Agents in Structured Finance review sample loan contracts against the data from the Loan Management System. IN-D now combines computer vision to convert image to a machine-readable document, NLP models to read, search and query the document; and then a similarity check model to review the content against a database.
WHAT IS BETTER?
Rapid Advancements in AI Means IN-D Products Have Monthly Releases
KYC – Fraud Detection
Additional checks based on computer vision have been introduced providing enhanced level of fraud detection for face morphing, face editing, text editing etc. Now for any ID document, Classification + Digitisation + Fraud Detection + Public Database Check are available through a single API call. Face Match and Signature Match modules have been enhanced. Some of the newer, enhanced models still require GPUs but we are working on making them more efficient to run on regular machines. Till then both options will be available.
Payables – Engineering Updates
In addition to improvement in the ML models for better accuracy, focus was on engineering aspects in the June release. The data extraction and validation modules are now independent for easy integration with RPA and low-code platforms. Dockerisation now allows for easy deployment across cloud environments. The front-end is now more intuitive for ease of review.
Onboarding – Beyond KYC
In addition to improvement in the ML models for better accuracy, focus was on engineering aspects in the June release. The data extraction and validation modules are Onboarding involves more than the KYC documents. With traditional software products, you will use a KYC software, may use one of the ‘bank statement analyser’ kind of tools and then process rest of the documents manually. For a company with strong focus on auto loans, IN-D is using the power of AI to classify all documents and route them to appropriate IN-D solution for processing. The whole process becomes completely digitized (with zero dropdowns).
New Partnerships and Updates on Current Partnerships
Digital Journey with Microsoft Power Automate
Power Automate is a service from Microsoft that helps you create automated workflows and rapidly rising up the RPA and Low Code automation charts. IN-D was among the first in the world to provide cognitive solutions as premium connectors for Power Automate. KYC, with a separate connector for India KYC; ICD10 Coding for Health Insurance, Invoice Processing etc. are some solutions from IN-D that you can easily ‘drag-and-drop’ to integrate with Power Automate. In most cases this extends the digital journey by 30-40% with just 5% extra spend.
IN-D KYC in Vietnam
IN-D has partnered with IDS for providing its Identity Verification solution Vietnam which used IN-D’s self learning model to train on different ID cards in Vietnam. Such machine learning based localization has ensured that IN-D will not only provide the best-in-class for fraud detection, but also be the most cost efficient option.