About
RPA and AI utilized as a smart assistant for audit tasks. The new process enables robots and artificial intelligence algorithms to collaborate auditor as a smart assistant leading auditor to focus on value added audit activities.
Innovation presentation
RPA and AI utilized as a smart assistant for audit tasks. The new process enables robots and artificial intelligence algorithms to collaborate auditor as a smart assistant leading auditor to focus on value added audit activities.
After the auditor defines the audit scope in terms of period and location for all loans, robot collects loan information from banking system and documents that received from customer during loan sale. Robot calls notifies A.I. to detect and interpret customer signature and validness of loan documents. A.I. responds and marks suspected loan documents.
As A.I. interprets and marks the loan as suspected, robot searches more data, pick related documents. Than it uses these whole inputs to organize audit material in a standard format. Robot prepares and files this material for auditor (human) inspection.
Auditor use this audit material delivered by fully automated process to examine loans based on alerts and focus on cases which have more risks in terms of audit objectives.
Smart automation of audit tasks initiative is on live for individual loan audit since June 2020.
Uniqueness of the project
Robots and A.I. collaborating and helping humans together especially in the audit domain is not ordinary and our project is pioneer in banking sector. A new audit environment is defined by human (auditor) – robot – A.I. collaboration. As the project is live, it brings auditor to have an understanding of getting benefit from smart assistant in non-value added audit operations. Auditor is focusing on value added risk assessment activities with a wider loan set.
Receiving assistance from A.I. comes with a new way of collaboration between auditor and smart algorithm. The auditors receive more and more effective support as they help in training A.I. In the project time, the auditors give feedback for 80,000 pages of loan documents in terms of signature existence and validness. This invaluable material has been used to retrain algorithm leading to improve its signature detection and verification capabilities.
The project changes the risk indicators to select cases for audit examination. New indicators cover similarity score created by smart algorithm for consumer signature on loan documents. Using smart algorithm to verify signatures on documents is a leading initiative in banking and has potential to be deployed to real-time document controls which can change transcation maker – checker mechanism in financial operations.
The other key aspect of the project is that smart text analytic and NLP algorithms support loan risk assesment process. The algorithms examine customer complaints and branch RM's issue tickets and give insights to auditors enabling risk assessment more effective.