Structured finance, the process of bundling assets like mortgages into tradable securities, is a cornerstone of modern finance. But the complexity of these deals demands rigorous due diligence—a process traditionally slow, costly, and prone to human error. Imagine sifting through mountains of loan applications and bank statements, verifying every detail. Now, imagine AI doing it for you. New research explores how artificial intelligence, specifically large language models (LLMs) like those powering ChatGPT, can automate this crucial verification process. By training AI agents on financial documents, researchers found they could accurately extract and cross-reference key information, like names, addresses, and balances, with impressive speed and accuracy. The study compared different LLMs, including open-source and commercial models like GPT-4, finding that while GPT-4 performed best, open-source alternatives offered a cost-effective solution. Interestingly, using two AI agents working in tandem—a “dual-agent” system—boosted accuracy even further. This points to a future where AI not only assists but potentially leads the due diligence process, minimizing errors and freeing up human analysts for more strategic tasks. While challenges remain, such as adapting to diverse document formats and navigating complex regulatory landscapes, this research highlights the transformative potential of AI in structured finance. As AI technology matures, expect to see even more sophisticated applications in finance, ultimately leading to more efficient and secure markets.