In many countries specially-abled individuals are requested with assistants to go in person to bank for physical identification and perform basic banking operation. In a world with much advanced technologies there is no unified solution help them to perform basic banking operation. The main objective for FinVA is to provide a platform where we can build banking application which are fully specially-abled friendly and create an inclusive environment. Voice interaction is a gateway to help improve personalization and the digital experience Voice technology has reached a stage where it started to disrupting the existing way of working. With technology companies applying for banking license and launching payment products they become clear competitors for traditional banking products. AI should be able to elevate customer experience.

80% consumers are satisfied with voice shopping experience 20% - 30% of mobile phone search are done by voice Voice shopping is expected to be around 10% of all mobile e-commerce. Make consumer experience convenient i.e) should not search for Sort code, Bill Payment due. It should be accessible by asking for it. It is not about answering basic customer support activities using chat bot instead doing all branch and customer transactional activities with voice. Preference towards voice technology is due to convenience and ability to engage hands free! Cost rationalization Reference: Pwc-voice-banking. • Free up front line branch staff • Reduce call center call volume • Minimize mid-/back-office operations

What it does

FINVA provides the platform based on ML & AI to enable voice assistance to any product. The user will be able to instruct the system using natural language to perform most of the operation. The transactions can be secured using additional authentication using any modes I.e) Face ID, Finger print, OTP or password. As a proof of concept for the platform few scenarios from customer retail banking are explained. Integrate next gen technology without much change to the core product.

How we built it

Started visualizing the concept based on one market demo. Built from scratch on iOS tech stack but can be easily moved to open tech stack like Python. Created a data based on the requirements of the FinVA. Converted the data to ML model and trained the model using python/CoreML, with accuracy of 80%.

Challenges we ran into

There are many challenges even small things like text to speech, speech to text, Natural language processing using ML and providing a chat like look and feel and many other challenges :-) Creating a seamless, user friendly chat like experience to consumers like their digital voice assistants ex: Siri/Alexa/Google

Accomplishments that I'm proud of

Using NLP to convert the natural language used by the user into a command specific to the application.

What I learned

Learning about ML the subset of AI Usage of NLP (natural language processing) Making sense of users language to a meaning full command and be able to execute it.

What's next for FinVA

FinVA is a platform using which the voice banking concept was explained. But the potential for this product is huge as it can be easily extended to other Essence suite of products and beyond. Voice technology will extend for both front and back office results in more efficiency which translate into competitive pricing for customer.

Built With

Share this project: