Last updated: 01 August 2020 06:42 AM
Added a few more challenges
Munch is built to lessen the barrier associated with the taboo of talking about money, lack of information around how much to save and invest, and the rather annoying and boring nature of financial planning activities. By leveraging the AA APIs, the app provides secure access and a holistic view into the user’s banking-related financial information, along with tools that will help the user optimize their spending habits. It starts with categorization via data enrichment to capture near-perfect spending patterns to identify areas that the user spends most on. Providing automated budget estimates, based on their historical patterns; predicting the end of month balance and spending projections (e.g., upcoming bill reminders, recurring spends, etc), and trade-off analysis (e.g., buying something extravagant, a weekend getaway, brand memberships, etc).
Understanding what the user will actually value. We carried out a survey with 230+ people along with 15+ user interviews to understand the pain points of the user associated with their finances.
Given the nascency of the ecosystem, integration with the account aggregators proved challenging, we co-ordinated with the great folks at onemoney to solve our issues one by one.
Security across our solution was and remains an area of continuous improvement. So far we've implemented biometric authentication and tight database security rules to never transfer any more data than is required.
We also spent a big chunk of our time and effort on transaction classification, here we managed to use a custom NLP model to read in the narration data within transactions and predict the general category along with the brand.
Another area of concern we had was the user education and onboarding journey, we managed to design a beautiful custom authentication flow and worked with our partner onemoney to deeply integrate with their sandbox to deliver a smooth, integrated experience.