Even today, many dread the idea of visiting banks, and going through the tedious policies in terms of documents and eligibility is a huge challenge while accessing loans. Banks have a stringent requirement in terms of high credit scores, an exhaustive list of documents to be produced and fulfilling the eligibility criteria.\n\n\n\nTo make these processes easy and fulfil every financial aspiration with the simplest, shortest and fastest way possible, veteran banker Manav Jeet founded Rubique. The sole aim behind Rubique is to identify gaps in the current lending value chain and deploy technology solutions to streamline and ease out the process challenges. And, they are driving it with a 15-membered team comprising of data architect, backend and frontend developers.\u00a0\n\n\n\nHow Rubique Makes It Easy\n\n\n\nRubique aggregates borrowers and lenders on its technology platform and provides a hassle-free experience to its borrowers through its matchmaking platform and end to end fulfilment model.\u00a0\n\n\n\nIn doing so it leverages machine learning and artificial intelligence along with big data analytics to build a technology system that matches the borrower with the right lender. \n\n\n\nRubique integrates credit policy of the lender on its system and uses self-training models through a feedback loop to provide custom offers to borrowers with high approval chances. Also, its Lending Gateway helps provide real-time decisioning and approval by integrating with the bank systems.\n\n\n\nAI Is Driving The Automated Marketing Solution\n\n\n\nAI is an integral part of Rubique. \u201cHaving spent 4 years in this space and working towards creating a sizeable base which is around 2,00,000 customers as on today, we have started work on deploying technologies like data science to build the cross-sell, upsell insights. We have also worked on Rubique Confidence score which makes use of alternate data & planning to market it to our FI partners to combine with traditional assessment parameters,\u201d says Jeet.\u00a0\n\n\n\nRubique Confidence Score relies on the digital footprint of the users to derive intent. Various sources of data that contribute to Rubique Confidence Score are SMS data, Device Data, Location Data, Call Data Record (CDR), Behavioral Data etc.\u00a0\n\n\n\nTheir in-house NLP-based SMS parser engine extracts more than 50 variables from SMS data per user. Some of the signals captured from SMS include salary info, employer info, running EMIs, credit card usage and more. Overall, more than 200 variables from various data sources are used to construct the Rubique Confidence Score.\n\n\n\nThe weightage of these variables is derived by using machine learning algorithms on training and testing data. Loan performance of various users serves as a feedback mechanism to improve the algorithm.\n\n\n\n\n\n\n\n\u201cWe are now working on rank ordering our score with Bureau score to define ranges for bucketing of customers.\u201d \u00a0shared Jeet.\n\n\n\n\n\n\n\nTechnology Stack At Rubique\n\n\n\nRubique flaunts a microservice-based architecture with each microservice dedicated for a particular task. As Jeet shares, they use REST API structures and hence integration with any system is very simple. They also use MEAN tech stack, Amazon cloud service provider, have polyglot data stores like MySQL, MongoDB, Redis etc. The data science algorithms are written in Python and Scala. Some of the other tools used are Hadoop & Spark, Redshift, Kafka and more. \n\n\n\nHow Is It Revolutionizing Finance Industry With AI-based Solutions\n\n\n\nDue to the varied risk appetite of financial institutions, the credit assessment & underwriting criteria are heterogeneous. The current market practices do not provide a decisive answer about whether the application will be accepted or rejected. \u201cDue to this \u201cMove in application\u201d mindset, the rejection ratio is in tune of 60%-70% leading to further credit gap which is essentially not because of the product unavailability but due to inefficiencies in right product discoverability,\u201d says Jeet.\n\n\n\nRubique is trying to address these challenges with below solutions: \n\n\n\nMatchMaking Engine: AI-based matchmaking engine is built in-house to provide custom-matched offers to the customer based on their profile, needs and interest. It takes into consideration various data points that a financial institution intends to capture. It then has a workflow engine which takes raw data input and derives a lot of other data points from it. It is then fed to the credit matrix and BRE to find the right set of offers. \n\n\n\nRecommendation Engine: AI-based Recommendation Engine takes into account the customer data captured for the loan\/card application and runs that data against the other credit policies of different products. \n\n\n\nGrowth Story \n\n\n\nRubique likes to differentiate itself from its competitors in its approach towards technology and thinking of financial services which have helped them launch various technology solutions for the industry. Rubique is unique in terms of providing a neutral platform where ender & borrowers can discover each other bringing transparency in the process, matchmaking algorithm and recommendation engine, technology-enabled distribution tapping every consumer segment, among others.\u00a0\n\n\n\nWith a portfolio of around INR 3,500+ crores loan disbursement facilitation through the platform, 150,000+ credit card setups & revenue generation of INR 70+ crores, Rubique is marching towards a leadership place in its domain. \n\n\n\nServing as a financial matchmaking platform for the entire ecosystem it is adopting to a physical approach to address the need of the entire ecosystem. \n\n\n\n\n\n\n\n \u201cThe next step for the startup aggressively strengthening the team by adding talent in areas of data science and products.\u201d says Jeet on a concluding note.