Automate loan intake with an AI-powered application
Project
Agentic Loan Intake Form
Year
2026
At The Mortgage Office, I helped define an AI-powered loan origination experience by conducting research, collaborating with stakeholders, designing workflows, and establishing a new design system while navigating AI, privacy, and legacy system constraints.
The Mortgage Office is an all-in-one loan origination and servicing platform used by private lenders to manage complex loan portfolios. While the platform is widely adopted, customers consistently reported usability challenges that created friction throughout the origination process.
Our goal was to create a modern borrower experience that leveraged AI to streamline document collection, reduce manual work, and simplify the loan application process.
Scope of Work
Problem
The existing origination process was manual, rigid, and difficult to navigate. Customer interviews surfaced:
Rigid intake forms that couldn't adapt to different loan products or lender workflows
Manual document review and data entry
Limited visibility into application progress and outstanding tasks
Constant borrower back-and-forth to collect missing information
Fragmented workflows pushing teams into email and external tools
The opportunity: Create a dynamic intake experience that lenders could customize to their workflows — while using AI to reduce manual work.
Solution
We designed an AI-powered borrower experience that simplifies loan applications and reduces manual effort.
Key improvements included:
AI-assisted document parsing
Dynamic and configurable forms by loan product and lender needs
Flexible, non-linear application flows
Borrower task tracking
Modernized design system
Scalable foundation for future AI features
Curating the Design System
Built from the ground up for the new AI-powered loan intake experience
Future-facing banking aesthetic using purple as the primary interaction color
Soft hues + rounded white surfaces across cards, tabs, modals, and sections
Creating reusable components for consistency and support future product growth
Phase 1
To reduce complexity, the first release focused on three loan types:
Fix & Flip
DSCR
Bridge Loans
The goal was to validate the borrower experience before expanding to additional products.
Process
Stakeholder Discovery
Partnered with Sales and Onboarding teams to understand workflows, identify pain points, and define opportunities for AI automation.
Customer Research
Interviewed 7 enterprise lenders to understand origination workflows and identify opportunities to reduce manual work.
Key Insight
Borrowers rarely complete loan applications in one sitting, leading to repeated follow-ups and delayed approvals.
This insight drove the design of a flexible application flow with AI-assisted document collection.
Challenges
AI Accuracy & Privacy
Worked closely with AI engineers to ensure document parsing was accurate, minimized hallucinations, and protected sensitive borrower information.
Flexible Workflows
Designed an application that allows borrowers to complete sections in any order without disrupting required loan dependencies.
Legacy Systems
Balanced a modern React-based experience with an existing servicing platform while working within engineering constraints.
