Case Study: How Citi and Numerated are working to transform lending technology through strategic partnership

Partnerships between established financial institutions and innovative startups are key to driving industry advancements. One such collaboration is the partnership and investment between Citi and Numerated, a fintech firm specializing in lending technology. In this exclusive interview, Katya Chupryna, Director at Citi, provides insights into the rigorous criteria Citi considers when evaluating potential fintech partnerships and investments. She emphasizes the importance of componentized solutions, integration capabilities, and an ecosystem approach to fintech investments.

On the other side of this partnership, Dan O’Malley, co-founder and CEO of Numerated, shares how their AI-powered platform has transformed Citi’s commercial underwriting process by reducing financial spreading and analysis time from days to hours. He also discusses upcoming enhancements to the platform that aim to further streamline operations and provide even more value to financial institutions.

This interview offers a behind-the-scenes look at the strategic collaboration between Citi and Numerated, highlighting the criteria that drive fintech investments and how technology is reshaping the future of lending. Whether you’re in the fintech space or simply interested in the intersection of technology and finance, this discussion provides valuable insights into what makes a successful partnership in today’s financial landscape.

What criteria does Citi consider when evaluating potential fintech partnerships or investments in the lending technology space?

Katya Chupryna, Director

The criteria for partnerships and investments in fintech companies, is not unique to lending technology space. Save for distinction between balance sheet intensive fintechs (which are lenders themselves and, therefore, require in many cases specific underwriting expertise, certain origination volume and performance track record), the criteria for pure technology providers includes multiple qualitative and quantitative topics ranging from management team’s experience to product-market fit and product roadmap to applicability across broader financial services industry. While the investment mandate is fairly open to fintechs of all stages, the investment strategy is an ecosystem approach—we are looking for companies who can work with each other and the overall industry to deliver the most impactful solutions. Similarly, there is a mix of horizontal and vertical solution providers in the portfolio depending on the needs of a specific business or enterprise-level opportunity.

Across all fintech verticals, we typically look for technology companies who provide componentized solutions, not just end-to-end platforms. The complexity of technological infrastructure at large financial institutions, paired with faster-than-ever development of external technology, leads a need for flexibility on vendors’ side to provide key pieces while also maintaining maximum flexibility to integrate with internally built or other vendors’ upstream and downstream systems. Instead of the previous decision model of buy vs build a monolithic system, we are now more often looking at joint development with multi-system integration capability and are prioritizing companies where such functionality exists. We are also seeing an increasing amount of partnerships among fintechs, where best solutions from various vendors are combined simplifying the integration lift on the end customer.

In case of Numerated, the solution is integrated with Citi’s home-grown wholesale lending technology stack. Numerated solution digitizes the financial spreading process, reducing processing time from days to hours and enables Citi to capture data from a wide range of sources and in various formats and build sophisticated analytics as well as structured data that can be upstreamed into other system.

Can you provide more details on how your AI technology reduces financial spreading and analysis time from days to hours?

Dan O’Malley co-founder and CEO, Numerated

Numerated’s platform revolutionizes the commercial underwriting process by transforming unstructured financial information into standardized, actionable data using advanced AI methodologies. Traditionally, this process is labor-intensive and error-prone, involving meticulous review of various document formats, manual data entry, cross-checking for accuracy, and performing calculations in individual spreadsheets. Underwriters then input this data into legacy systems for credit risk assessments, a journey that can take days or even weeks.

Our platform automates and streamlines this entire process. By leveraging Optical Character Recognition (OCR) and Natural Language Processing (NLP), we reduce manual data entry from hours to seconds while maintaining high accuracy. Extracted data is standardized to bank-specific models through multiple AI layers, including large neural networks and logistic regressions, enhanced by user feedback.

This AI-driven approach converts complex data into actionable insights for lending decisions, reducing analysis time from days to minutes. Our continuous innovation, including the introduction of large language models, further enhances the power of technology and data within the industry, making the underwriting process more efficient, accurate, and cost-effective for banks.

How does Numerated ensure data accuracy and consistency when processing diverse financial statement formats, including handwritten notes?

Numerated ensures data accuracy and consistency when processing diverse financial statement formats, including handwritten notes, through a combination of advanced AI technologies and methodologies:

1. Optical Character Recognition (OCR) 

Our platform uses OCR to accurately digitize printed and handwritten text from various document formats, transforming them into machine-readable data.

2. Natural Language Processing (NLP)

NLP processes and interprets the digitized text, ensuring that context and meaning are preserved, which is crucial for understanding financial statements accurately.

3. AI Layers for Standardization

Once data is extracted, multiple levels of AI, including large neural networks and logistic regressions, standardize the data to bank-specific underwriting models. This process ensures that diverse financial formats are consistently translated into a standardized format.

4. User Feedback Loops

Continuous user feedback is integrated into our system, allowing the AI to learn and improve over time, enhancing the accuracy of data processing and ensuring that any anomalies are quickly identified and corrected.

5. Accuracy Checks

Built-in accuracy checks throughout the process help maintain a high level of precision. These checks cross-verify the extracted data to ensure its consistency with the original documents.

By leveraging these technologies, Numerated can process diverse financial statement formats with high accuracy and consistency, turning complex, unstructured data into reliable and actionable insights for lending decisions.

Looking ahead, what new developments or enhancements to your platform are you planning to implement?

Numerated plans to implement several new developments and enhancements to further elevate our platform’s capabilities:

1. Generative AI for Predictive Analysis and Narratives

We are integrating generative AI to enhance predictive analysis, providing comprehensive summaries and insights based on existing trends. This will improve the accuracy of future financial outcome predictions and offer a deeper understanding of potential risks and opportunities.

2. Increased Automation and Processing Time Reduction

Our goal is to continue to increase our processing speed and increase the level of automation on our platform. We are continually working to reduce processing times and increase efficiency, allowing underwriters to focus on higher-value tasks.

3. Third-Party Data Benchmarking

We are incorporating third-party data sets for robust benchmarking and analysis. This will provide  underwriters broader context and enhance the quality of credit assessment by enabling comparisons against industry standards and trends.

These enhancements demonstrate our commitment to leveraging cutting-edge AI technology to improve the efficiency, accuracy, and depth of financial analysis for our users.