Designing the end-to-end borrower experience for AI-powered digital mortgage platform — from pre-qualification through to closing — transforming one of the most stressful financial journeys into a guided, transparent process.
Mortgage AI had built powerful AI technology capable of automating underwriting and dramatically accelerating mortgage approvals. But their borrower app was haemorrhaging users: only 28% of applicants who started the application completed it.
The existing app was a digitised paper form — a 45-field application with no guidance, no progress indication, no clear explanation of what documents were needed and why, and no AI personality whatsoever. Users felt interrogated, not guided.
"I got halfway through and had no idea what I was agreeing to, why they needed my last two years of tax returns, or how long this was going to take. I just gave up."
— Borrower, Exit SurveyNo progress indicator, no estimated timeline, no explanation of what happens with submitted data. Borrowers felt they were filling a black hole.
Users didn't know which documents were needed, in what format, or why. Incorrect uploads caused 40% of application delays.
AI capability was completely hidden. Users experienced the same wait times as traditional lenders — the competitive advantage was wasted.
I ran a 4-week research sprint combining moderated usability sessions on the existing app (12 participants), diary studies tracking borrowers through their entire mortgage journey (8 participants over 3 weeks), and analysis of Hotjar session recordings to identify precise drop-off points.
The diary study was particularly revealing — it showed that mortgage anxiety peaked not during the application itself, but in the silence after submission. Borrowers didn't know what was happening, who had their documents, or when to expect next steps.
Step 3 (income verification) was the #1 drop-off point — 34% abandoned here
Document upload section triggered 67% of support tickets
Borrowers wanted real-time status updates more than faster processing
Plain English explanations of mortgage terms reduced abandonment by 24% in A/B tests
25-35 years old, buying their first home. High anxiety, low mortgage literacy. Needs guidance at every step, plain English explanations, and reassurance that things are progressing.
35-55 years old, refinancing an existing mortgage. Has been through the process before. Needs speed and efficiency — hates re-entering data they've submitted before.
Has complex income documentation. Standard mortgage apps weren't built for them. Needs clear guidance on which documents apply to their situation.
End-to-end borrower journey map across 14 touchpoints from search to keys
Usability sessions, diary studies, Hotjar analysis, 200+ exit survey responses
2-day design sprint. Three concepts explored: wizard, chatbot, and task board
100+ screens across iOS, Android, and Web. Full interactive prototype for testing
5 rounds of usability testing. A/B tested 3 critical flows in staging environment
Weekly Hotjar review, monthly cohort analysis, continuous iteration backlog
A persistent "Your Mortgage Journey" sidebar shows exactly where the borrower is across 6 stages — Pre-Qual, Application, Document Review, Underwriting, Approval, Closing. Always-visible, always accurate.
Instead of a generic upload button, an AI document coach asks 3 questions about the borrower's employment type and income sources, then shows exactly which documents are needed — with sample images and format guidance.
The silence after submission was replaced with a live AI processing feed — "Your W2s have been verified ✓", "Income confirmed ✓", "Debt-to-income calculated ✓". Borrowers feel in control, not in limbo.
The redesigned borrower app launched in Q4 2023. Within 3 months, application completion had jumped from 28% to 73% — a transformation driven by reducing anxiety, improving guidance, and making Mortgage AI advantage visible to borrowers.
Application completion rate jumped from 28% to 73% — a 2.6× improvement
50% reduction in processing time through AI document coaching reducing incorrect uploads
$2.1M additional year-1 revenue from incremental closed loans
App Store rating climbed from 2.9★ to 4.7★ — driven by "easy to use" reviews
67% reduction in document-related support tickets
Post-closing NPS of +68 — up from +12 on the previous product
"I've done two mortgages before. This was the first time I felt like the process was actually on my side. The AI status updates were incredibly reassuring — I always knew exactly what was happening."
— Borrower, Post-Closing NPS SurveyThe core insight: anxiety is the enemy of completion. Most UX work on mortgage apps focuses on reducing steps. But steps weren't the problem — uncertainty was. Borrowers didn't abandon because it was too long; they abandoned because they didn't know what was happening or why it mattered. Designing for transparency and reassurance — not just efficiency — was the real breakthrough.
I'm open to senior UX roles and consulting engagements.