CXO of the week: Mathangi Sri Chief Data Officer, Yubi

Yubi, formerly known as CredAvenue, is an Indian fintech company that connects businesses with financial institutions, banks and other lenders through its digital platform. Yubi is re-defining global debt markets by freeing the flow of finance between borrowers, lenders, and investors.

They started the journey in 2020 with a simple goal – to deepen debt markets and truly unlock the potential of Indian enterprises, that had thus far, been inhibited by the inadequate access to capital. This vision gave life to a revolutionary innovation in the debt market called CredAvenue. At inception, CredAvenue catered to different types of capital requirements: securitisation, bonds, co-lending, supply chain, and loans.

However, as it grew in scale, size, and mission, they knew that there needed to be something bigger to encapsulate our way forward. That is when Yubi was born.

Mathangi Sri Chief Data Officer, Yubi. With over 107 global patents publishes and 10 global patent grants in the area of data science and machine learning, Mathangi has been at the helm of driving growth for Yubi. 

Recently we have engaged in an interview with Mathangi Sri, Chief Data Officer, Yubi. She shared an overview of how Yubi has adopted AI/ ML into its product offerings, her entrepreneurial journey, the company’s growth, and much more.

Introduction.

In the span of five years, India has made great strides towards its goal of becoming a $5 Trillion economy. Through initiatives such as Digital India and Start-up India, the government has provided a platform for entrepreneurs to thrive. Similarly, the introduction of UPI has enabled India to become a pioneer of the future of payments. Despite these advancements, the country's lending infrastructure is still limited in comparison to the digital revolution the world is going through. This is especially true for the 8 million MSMEs in India, whose potential for growth is hindered by inadequate access to credit. To address this issue, we at Yubi (formerly CredAvenue) are working to create a unified credit infrastructure to promote transparency and efficiency in the lending ecosystem.

Yubi's product suite covers business loans, co-lending, fixed income securities, supply chain financing, asset-backed securitisation, real estate and infrastructure financing. This allows lenders to find promising enterprises, execute deals and disburse, manage and monitor capital digitally through a single platform. With the acquisition of spocto and Corpository, we also have collection capabilities and have enabled lenders to monitor their portfolio performance, thereby increasing trust in the largely trust-deficit economy and enabling credit to reach more creditworthy enterprises. Essentially, we cover the end-to-end debt lifecycle under a single platform.

A brief overview of how Yubi has adopted AI/ ML into its product offerings.

We have adopted AI and ML into our product offerings to enhance our data-technology capabilities and to provide a better customer experience for borrowers and investors. We use ML models to power decisions across the entire credit life cycle journey of millions of transactions, including recommendation models, credit risk models, collection propensity models, NLP engines, image detection algorithms, etc. We also use AI to power use cases across products such as fraud & risk scorecards, deals recommendations, better investor discovery, next best action recommendations for relationship managers, channel engagement strategy for collections, computer visions for document validations, and auto commentary through natural language processing. To support these initiatives, a dedicated team builds infrastructure for deploying real-time and batch models for real-time integration. This infrastructure has allowed us to reduce our time to go to market for the ML models and has accelerated business impacts. In addition, we also provide powerful dashboards and data access controls for non-data native users to enable data-driven decision making.

How can small financial players overcome the challenges faced while adopting technology into their businesses?

Up until now, traditional lending institutions and small players have been resistant towards the debt market due to the high cost of digitizing and training staff. However, we are enabling borrowers to connect with them, reducing paperwork, and allowing data-driven decision-making, thus improving credit decisioning efficiency while decreasing operating costs and turn around times, allowing for fair and transparent access. Additionally, when we drive digital transformation for lenders, it can be implemented all at once across different geographical areas, making the scale not an issue, but rather an opportunity.

Yubi not only provides access to a network of lenders and borrowers across the country, but also helps manage debt workflow, decisioning and monitoring risk factors and portfolio performance. As an example, Yubi Co.Lend has allowed banks to integrate with multiple partners with one-time API integration, and in one month’s time. This has led to the disbursement of INR 10,000 Crores on our platform.

How is Yubi driving innovation in the fintech sector with emerging technology solutions; an overview of YubiBERT?

Natural Language Processing (NLP) has been a crucial part of many tech companies and their success. However, we noticed three main pain points. Firstly, India being a complex market with multiple languages, there was no model to analyze regional languages. Secondly, domain specific models perform better than generalized state of the art models. And finally, despite being the world’s biggest and the most innovative fintech market, Indian fintech companies are compelled to use large language models (LLMs) which are not designed for the fintech sector or the Indian context. While there are domain specific models for fintech, none of these models consider the vastly different context of the Indian fintech market. This has resulted in multiple inefficiencies for the fintech sector.

With YubiBERT, we aim to solve these problems for the entire fintech industry so that the ecosystem can collectively grow. YubiBERT was trained with 200+Gb fintech public data and over 1 billion sentences, making it one of the most robustly trained language models in the world. When fine tuned on FinTech related NLP tasks, it performs better than BERT, RoBERTa, FinBERT and DistilBERT models. We are positive that this will have a massive impact on the fintech community and we are excited to see how the data science community takes this language model to the next level.

Outlook for tech innovations in 2023.

Data and AI governance will take centre stage: With the Union Budget announcement of the National Data Governance Policy this year, it will enable access to anonymised data. Additionally, the government is setting up three centres of excellence for Artificial Intelligence in top educational institutions. This will allow industry players to partner in research, development and problem-solving. Moreover, it will ultimately, result in a strong AI ecosystem and the nurturing of quality human resources in the field.

Explainable deep learning models will power credit decisioning: They enable lenders to gain a better understanding of the factors that drive a credit decision without compromising on the prediction performance. By understanding the underlying logic behind the decision, lenders are better able to adjust their credit policies to be more effective and fair. Additionally, the transparency provided by explainable models helps to build trust between lenders and borrowers, improving the overall customer experience.

Compliance on loan collection interactions - voice calls, chats will be completely led by AI & ML solutions: without compromising on the prediction performance: Collection conversations are generally prone to being harsh and abusive. Hence, we expect regulation to play a critical role in monitoring customer interactions during collections. Speech recognition with state of the art language models would be needed to monitor the abuse and harshness detection in a collection interactions.

Why do you think there is a need for your company to exist in today's market?

Gaining access to working capital in the Indian market is becoming increasingly significant and its availability could also have an impact on the economy on the whole. Historically, India has suffered from a lack of credit because the market is not very populated with participants. Recent data has shown that an estimated 3% of entities are controlling nearly 50% to 60% of the overall outstanding debt in India, leading to a credit gap for the middle market, or those triple B rated companies. The Indian government has been trying to stimulate capital market participation, yet it remains relatively shallow. Yubi in its drive for greater financial inclusion is providing a digital solution to connect all participants, enabling borrowers to access lenders and investors all over the country with a single click. Additionally, the platform offers clients a customized dashboard with a clear view of the process and 24/7 reporting support. Moreover, Yubi’s platform offers users a number of debt instruments, such as bonds, term loans, and supply chain financing, through its seamless interoperability.