Fintech Infrastructures 🏗️

Fintech products have grown dramatically in the last couple of years and will evolve even more in the next few years.

One main USP of most fintech apps in COnveninve is better service and UX.

For Banks or NBFCs, user experience is not a major KPI.

They work on basic baking stuff like onboarding, payments and more.

They also have to protect user privacy and data to ensure the safety of innovation and growth.

But when Tech companies started reaching out to these banks, they had to work operationally and technically differently.

Now their customer’s data is not going to be in their investment, it has to be shared as well.

It means they have to work on both inward and outward customer data differently.

To do that, depending upon the required features, banks and fintechs work on different types of system design.

I’ll try to explain a few of them below:-

Version 1 :- Direct Integration

Consider this framework when you want to buy insurance from a fintech app.

You enter PAN details, address and more on the app, then the app’s server fetches all relevant information from the insurance company’s backend via APIs.

In the case of initial neobanks, onboarding and complete transaction and card management used to work like this.

Good part:-

1. Quick setup
2. Access to consumer data
3. Personalized offers and engagement
4. Easy to service

Bad part:-

Fintechs have access to PII where they onboard people who are not their customers. (maybe you have completed the onboarding on NeoBank but you are the bank’s customer, not NeoBank’s)

Version 2 :- Partial SDK method

This practice is introduced to protect some PII and not let fintechs have access to such information.

Such as Aadhaar validation, entering m-PIN, video KYC flow and more.

Example:-

Step 1: The user clicks on the ‘Validate Aadhaar’ CTA in the app’s frontend.

Step 2: AThe pp’s server-side calls bank SDK’s ‘Validate_Aadhaar’ function

Step 3: A UI is shown to the user which blends seamlessly in fintech’s UI, so user experience doesn’t get compromised.

Step 4: The user successfully validates everything and SDK gives a successful response to the app server side.

Step 5: The user sees a success message in the app’s frontend.

This SDK approach is not limited to only financial products. Whereever, users have to expose their PII, and the business does not have permission to store such information except to validate the user’s existence.

Apart from that storing and processing Aadhaar data requires businesses to have an Aadhaar Vault.

The good:-

1. Fintechs can still consume user activity data to provide them with personalized experience.
2. User’s PII is not getting stored at fintech, as they are not fintech’s customers but the Bank’s.

The bad:-

1. As fintechs do not hold customer PII, targeting them becomes difficult.
2. They hold activity to build an engine but it is disassociated with PII.

Version 3 :- Full SDK approach

Now, this is the safest approach to manage PII while making sure user’s experience doesn’t get compromised.

Here app’s server side is also hosted in the bank’s infra.

Step 1:- Beta version is built and deployed in fintech’s infra.

Step 2:- Bank audit the build and complete setup

Step 3: The whole codebase gets moved to the bank’s infra and credentials are shared with the fintech dev team to mange the code.

Step 4:- Depending upon the business agreement, specific SDK functions and APIs get exposed to the fintech server side.

Here everything is the same as V2, the only difference is that, fintech do not have access to plain customer activity data.

Everything is encrypted and very few people have access to manage the code, analyze data and service the customer.

In this case, building a plain dashboard becomes a challenge, so let’s not talk about a data model here.

Personalization becomes pure activity and behavior-based.

You can practically never target customers directly, unless they reach out to you for something.

Move forward

Now, these are some examples of where I have worked, certainly, there must be safer and more complicated model to work on.

Each comes with its own, pros and cons and mostly a business decision on which model they can go forward with.

As I have said before, fintech works on a very thin line of what users want and what regulations allow.