Cover Image

Machine Learning, but with some Humanity

May 17, 2021 - Reading time: 5 minutes

In our previous article we explored machine learning as a branch of artificial intelligence. In this article we will now look at how Intergreatme has deployed this kind of technology in our business and consider how it enables our products in light of the domestic regulatory environment.

Intergreatme has a responsibility for the security and upkeep of our platform around the clock. To accomplish this Integreatme has automated our document processing in a way that provides scalability to process large transaction volumes by embracing machine learning.

The machine-learning document processing occurs after we have collected key pieces of information. This is then coupled with information from what we call ‘golden source’ databases, like the Department of Home Affairs HANIS application programming interface (API).

When a transaction begins, we queue data which allows us to monitor the user journey from start to finish. This is done in order to identify transactions which require assistance, at which point they can automatically be directed to our human verification operations team to provide the necessary support which can be directed either to the user, or to our customers via our innovative Customer Insight Platform (CIP). At every step of the verification process, documents are inspected in real-time, resulting in near instant validation, verification, or feedback as the user uploads their supporting documentation.

The back office verification team also inspects transactions, ensuring that we flag transactions that are potentially fraudulent. What this means is that when our machine learning system cannot make a good decision, it refers it to a human being in order to make a final determination. This ensures maximum efficiency and compliance, and that our systems operate in the most balanced way possible.

#intergreatme #remotekyc #kyc # digitaltransformation #onboarding #knowyourcustomer #regtech #fica #rica

Categories