Traditional underwriting is slow, subjective, and full of paper—but it doesn’t have to be. Learn how smart lenders are streamlining credit decisions and staying competitive in equipment finance.
In this Q&A, Monitor Editor-in-Chief Rita Garwood speaks with Carolina Patiño, Product Manager at Kin Analytics, about how data and technology are reshaping the underwriting process in equipment finance. From predictive models to OCR and alternative data, Carolina explains how lenders are upgrading legacy workflows, reducing risk, and delivering faster, smarter credit decisions without sacrificing human judgment. Listen to the full podcast here.
Rita Garwood: To kick things off, how would you describe the traditional underwriting process in equipment finance—and what are its biggest limitations?
Carolina Patiño: Underwriting includes multiple complex steps: application intake, credit analysis, and decisioning. But three key issues stand out. First, the process is extremely manual—entire teams spend time keying in data from documents such as bank statements or IDs, which is tedious and error-prone. Second, there’s subjectivity. One underwriter’s red flag might not be a concern for another. Third, there’s a lack of accessible, centralized data. Analysts often can’t get a full view of repeat applicants or industry conditions. All of this slows down decision-making and leads to less optimal credit outcomes.
Garwood: What’s pushing lenders to move away from intuition-based underwriting toward data-driven models?
Patiño: Speed and competition. Borrowers today expect decisions in hours, not days. Lenders need to move fast while still managing risk. Automation and data analytics give them that edge. Also, companies are starting to realize the untapped potential of the data they already have—and how it can improve accuracy and efficiency across the board.
Garwood: What tools or technologies are making the biggest difference in underwriting right now?
Patiño: We’re seeing huge progress with OCR (Optical Character Recognition) and LLMs (Large Language Models). They’re transforming the document-heavy intake process by extracting data from messy PDFs, even handwritten forms, and automatically creating deal entries in CRMs. But predictive analytics is the real game-changer. Lenders are moving away from relying only on third-party bureau scores and toward building custom models using their own portfolio data. These models reflect real customer behavior and enable more tailored, accurate decisions. AI is also powering chatbots, document validation, and due diligence support.
Garwood: Alternative data is getting a lot of attention. What non-traditional data sources are actually working in credit models today?
Patiño: Industry-specific data is especially valuable. In transportation, for instance, we’ve used the DOT’s Safer database, which includes fleet size, crash history, and compliance data. We’ve also worked with GPS data to detect risk, like if a vehicle is in a repair shop for weeks—it can signal financial stress. We’re even using satellite imagery and digital footprints to verify businesses. These alternative sources provide more context and early warning signs that traditional data cannot.
Garwood: Despite advances in tech, manual processes are still common. Where are they most persistent, and why?
Patiño: The biggest choke point is still in the application intake stage. It’s highly manual, with varied document formats and a lot of back-and-forth for missing info. One reason automation hasn’t fully taken hold is complexity — no standard formats mean tech has to account for a lot of variation. There’s also distrust. Even when OCR tools are accurate most of the time, a single mistake can cause people to reject the entire solution. Finally, there’s the cost and resistance to change in a legacy-driven industry. Change management is a big part of the challenge.
Garwood: What’s the best way for lenders to start modernizing these processes?
Patiño: Start by mapping every step of your underwriting process. Identify where delays happen—whether it’s data entry, communication with brokers, or background checks. Then focus on high-volume, low-complexity problems first. For example, automating application forms before tackling complex legal documents. The most successful lenders also adopt hybrid solutions: using AI to handle the repetitive tasks and leaving complex analysis to experienced underwriters. And crucially, involve your teams in the process so you’re not just imposing change from the top.
Patiño How do you strike a balance between innovation and regulatory compliance when deploying new credit models?
Carolina: First, use clean, non-discriminatory data. Don’t include race or gender in modeling. Second, ensure explainability. Blend high-performance AI models with transparent models when building scorecards so you can clearly justify decisions. And finally, document everything—your data sources, methodologies, and how you’re using the models. At KIN, we’re GDPR and SOC 2 compliant, so our clients can be confident that their data is safe and their modeling is defensible.
Garwood: What real-world results have you seen from data-driven underwriting?
Patiño: We’ve seen default rates drop by about 20% and approval rates rise by around 30% for clients using our scorecard solutions. One of the biggest wins is turnaround time — applications can be scored in seconds, giving analysts more insight to make faster decisions. In early testing of our application submission automation product, we’re seeing deal entry time cut by 75% — from the moment a deal hits your inbox to it being live in your CRM.
Garwood: How does faster, smarter underwriting affect the borrower experience—especially for small or non-traditional applicants?
Patiño: It’s huge. Borrowers want to get their equipment and start generating revenue as soon as possible. Even small gains in efficiency make a big difference in their experience. Faster decisions, fewer delays, and more tailored assessments help build trust and get deals funded faster.
Garwood: For lenders trying to scale without overwhelming their credit teams, what’s the most effective first step?
Patiño: In addition to process mapping, focus on your data infrastructure. Start collecting and organizing your data now—even if you’re not yet sure how you’ll use it. Clean, well-structured data is the foundation for any AI model. And look for quick wins: automating repetitive, low-risk tasks to free up analysts for more critical thinking.
Garwood: What mindset shift do underwriting leaders need to make today?
Patiño: Stop viewing tech as a threat to human judgment — it’s a tool to enhance it. The goal isn’t to replace underwriters but to empower them. Underwriting should be designed to say yes to the right deals — and that requires trusting data over gut feeling, and making tech a partner in smarter, faster decision-making.
