top of page

BIC Pledging Review

An Aurora Document Review Application designed to radically transform the Document Review Workflow with its UX and LLM capabilities.

This page provides a high-level overview of my role and the outcomes delivered; detailed problem framing, process, and design decisions rationale are intentionally reserved for in-depth discussion.

About the Project

Doc Review, BIC and CBLA reviews loan documents, indexes, captures metadata and pledges loan to FRBNY / FHLB during the life of the loan. This is a manual process and time-consuming due to deal complexity. A very high 180UPT / doc review case (Unstructured Docs) incl 45 mins for BIC pledging leading to huge backlog and multiple rounds of QC. 

​

378K+ documents indexed annually | 20K+ checklists | 500K+ data elements captured | 11,029 annual hours | 9 systems to complete 1 Aurora Review

The Challenge

Read and Interpret 11 document types to answer 43 BIC/Pledging questions leveraging LLM capabilities; enabling loan processors to complete BIC/pledging eligibility review in minutes

My Role

  • Plan and conduct user research, benchmarking existing solutions and processes

  • Create personas, map end-to-end task flow

  • Determine improvement areas and opportunity to leverage LLM/AI

  • Translate concepts into user flows, wireframes, mockups and prototypes

  • Conduct usability testing, and push adoption by monitoring user feedback

  • Helped in data labeling exercise to determine effectiveness of AI

My Approach 

  • Meet with stakeholders to clarify business objectives, user pain points, desired outcomes, project goals, success metrics, and constraints. 

  • Conduct interviews and shadowing to uncover user needs, behaviors, and pain points; analyze environments and workflows for real-world challenges. 

  • Evaluate internal and external solutions to understand industry standards and innovative approaches. 

  • Map current and ideal user journeys, highlighting steps, decision points, bottlenecks, and opportunities for AI/LLM integration. 

  • Work with data scientists and engineers to identify where LLM/AI can automate, assist, or personalize; prioritize improvements by user impact, feasibility, and business value. 

  • Facilitate brainstorming, sketch concepts, and validate ideas with stakeholders and users for early feedback. Mockups 

  • Create detailed user flows, wireframes, and high-fidelity mockups, ensuring intuitive, and AI-integrated designs. 

  • Build interactive prototypes, conduct usability testing with representative users, and refine designs based on feedback. 

  • Support data labeling for AI model training, collaborate with data teams to measure effectiveness, and inform design iterations. 

  • Roll out the solution, monitor adoption, collect feedback, and iterate post -launch to optimize user experience as AI evolves. 

The Result 

  • Lending operations process optimization and improved efficiency by 80%

  • Opportunity to review and close approx. 4Bn $ outstanding loans / month (in prod)

  • Validation of LLM Q&A capability

bottom of page