Sherlock Homes

Recent loan applications dashboard
Fun fact: the Sherlock Holmes IP went public domain in 2023. Our name is fully legal
Analysis engine for the quick pre-qualification feature
teams can view data from all previous loan applications in one convenient location.
A low-risk applicant. Highest risk score is 100, applicant scored 10. AI recognizes low holistic risk in spite of a high loan-value ratio.
AI report for the same application; applicant scored 0% (perfect score) on all social metrics indicating low lifestyle risks.

Inspiration

Inspired by challenges faced by family members in the real estate industry, we created Sherlock Homes. Risk assessment is a lot like detective work, but it's currently slow, manual, and prone to human error. We saw the need for an AI-powered investigator that replaces tedious data-entry with high-speed, intelligent risk analysis.

What it does

Sherlock Homes streamlines the underwriting process by automating repetitive tasks and extracting meaningful insights from social media and public records. It highlights key risk indicators while identifying oft-overlooked factors, helping underwriters make faster, more informed decisions with greater confidence.

How we built it

We used Claude Code to assist in developing backend Python scripts that integrate multiple AI systems, including Perplexity, OpenAI, and keyword analysis models, all made possible by Keywords AI’s API. These tools work together to process large datasets, interpret sentiment, and generate actionable reports for underwriters.

Challenges we ran into

Integrating real-time social media sentiment data into the frontend was a major challenge, especially when dealing with API restrictions and limited access to platforms like LinkedIn. We also faced debugging hurdles when ensuring compatibility between various AI models and maintaining data accuracy.

Accomplishments that we're proud of

We’re proud to have successfully debugged and optimized our financial analysis AI, ensuring reliable outputs and meaningful insights. Overcoming integration challenges and achieving a functional prototype that connects multiple AI systems stands out as a key milestone for our team.

What we learned

Through this project, we gained deeper insight into the real estate underwriting process and strengthened our technical skills in data analysis, API integration, and AI model coordination. Most importantly, we learned to investigate gaps in pre-existing industries where AI is an untapped productivity multiplier.

What's next for Sherlock Homes

We would love to offer an extension for banking services so businesses can manage everything from their own website. We plan to scale by attracting investors with connections in the banking industry to whom we can sell our product.