Developing a Fintech Chatbot

In fintech, there are often billions of data points, enclosed in systems that can only be accessed by the few. We helped Quant Insight to take those data points and use them to power a chatbot that made the information available to many.

The challenge

Quant Insight produce macro economic views for analysts within financial institutions. Their financial models include billions of data points, but they’re all hidden away in complex software that can only be accessed by data scientists. Just think - all of that really useful information, that retail investors would love to have but couldn’t yet access.


How AI helped

QI worked with us to develop a retail investor chatbot that could query their financial models using Natural Language Processing (NLP). We developed an intent engine that took a natural language question (intent) about a stock or market and provided an answer with a rating for that stock.

Most retail investors don’t have the time or competence to access complex algorithms, so these answers were built into a natural language chat interface - otherwise known as a chatbot.

The bespoke NLP work Deeper Insights™ (formerly Skim Technologies) carried out was to identify the factors affecting the stock, and return a relevant answer. Comprehending thousands of different example questions for hundreds of different companies and stocks, to find the right one.

The result

QI’s retail investors were able to easily access stock information via a friendly chatbot. The chatbot was able to achieve over 94% accuracy for selecting the correct stock and factors related to a question. An impressive outcome.

“Deeper Insights™ (formerly Skim Technologies) were a pleasure to work with, and extremely knowledgeable in the field of NLP. Their ability to take our idea and turn it into a working product made them the perfect partner for our fast-growing business.”

Mahmood Noorani, CEO, Quant-Insight