P

Prevent consent fraud and Portability to new AA

Leverage AI/ML models to prevent consent frauds related to ‘Type of data’ and ‘How much data’ asked by FIU. Also, avoid vendor lock-in by portability of account and consent to a new AA

Carousel Gallery Item: 1
Carousel Gallery Item: 2
Carousel Gallery Item: 3
Carousel Gallery Item: 4
Carousel Gallery Item: 5

Last updated: 31 July 2020 04:48 PM

Created project

The problem Prevent consent fraud and Portability to new AA solves

Our solution solves key aspects of consent frauds and also proposes a way to port customer account and consents to a new AA.

With the demographic diversity in India, many AA ecosystem users may not be financially literate to easily understand ‘What type of’ and ‘How much’ data is asked by a service provider or FIU. To enhance customer experience better, prevent consent frauds and gain customer confidence in this new data sharing ecosystem, we propose Value added services that AA could offer to their customers.
We propose solution where AA could consume consent requests and build AI/ML models that will learn overtime and flash flags based ‘What type of’ and ‘How much’ data FIU is asking:
a) Module 1: Data driven alerts (How much): Additional data than industry standard for a given purpose
b) Module 2: Sensitive score (What type of): Data sensitivity is high considering the type and method the data is asked by FIU
Additionally, to avoid vendor lock-in, we propose high level API thinking for AA portability. This could be achieved with seamless customer experience and bring level playing field for all AAs
c) Module 3: AA portability: Customer should be able to port Account and Consent easily to a new AA

Challenges we ran into

A. Business Challenges:
• Creation of sample Industry standard data required corresponding to each purposes (Like personal loan, Deposit, SIP etc.)
• No clarity on central repository's functionality and integrations
B. Technical Challenges:
• ML Modeling with sample data for Sensitivity score and Data driven alerts
• Finalizing the algorithm for the model analysis
• Optimizing user experience and integration with sandbox

Discussion