back to notes

MLCI (Machine Learning Customer Insights, DDA & Credit Cards) Minimal Viable Experiment – POC Effort

MLCI (Machine Learning Customer Insights, DDA & Credit Cards) Minimal Viable Experiment – POC Effort

Description: Machine Learning use cases to analyze customer financial data and provide insights on financial patterns
Why Labs?: Labs provides both expertise and infrastructure to conduct a POC within the timeframe of this effort.
Key Question: Can machine learning algorithms be developed based on existing DDA and Credit Card information and be used to effectively predict customer financial patterns?
How POC learnings will be Utilized: If the POC is successful, efforts for production implementation can be undertaken

Scope and Duration: The scope of the effort includes the following use cases
Review your Balance: Create a model to predict that current balance (including expected deposits) for account may be too low to cover upcoming anticipated spending. Data: DDA
Provider bill is higher than usual: Create a model to predict that latest payment to a service provider is higher than usual. Data: DDA
Expected Deposit on Date: Create model that a regular deposit (salary, social security, etc.) has not arrived on expected date. Data: DDA
Recommendation to set-up an AutoPay from Checking Account to Savings Account: Create a model to recommend use of recurring transfer from checking to savings account based on payment patterns. Data: DDA
Suggest Person-to-person payments (P2P): Create a model to recommend Person-to-person payments (P2P) based on previous payment patterns. Data: DDA
“Need to Watch” notification - Auto Renewal: Create a model to identify a purchase with a merchant that is flagged as a free trial merchant, after free trial expiration or an automatic renewal process. Data: Credit Card
Out of Scope:
Notifying users (This is only a POC: no customer is involved)

Strategic Alignment/Fit:
Predictive Banking