Automate critical processes in retail banking

Retail Banking

Many data collection tools such as forms are not optimised for data analysis. In today’s data driven world, it’s vital that models are able to ingest all available data in order to maximise performance.

Problem 

A large financial institution had many examples of forms that had been filled in by individuals, largely with free text data. Their existing rule based solutions had failed at classifying this information, and the client was looking to expedite analysis without relying on outsourced, mechanical turk style services.

Solution

Mind Foundry’s in-house expertise on Natural Language Processing proved invaluable in this task, as we were able to analyse the free text data rapidly and efficiently.

In less than two days using the Mind Foundry platform, a data scientist working for the client had created a prototype solution that could directly ingest raw data, parse and process it to make predictions and send these back to external systems.

By analysing the free text data rather than structured data fields alone, performance was significantly increased, making the machine learning solution a viable or even improved alternative to the existing human-driven solutions.

Results