## Dwella – AI-Driven Mortgage Automation Platform

##### Industry

##### Social media

##### Project type

##### Case study

##### Date

###### March 7th, 2024

### Project details

**Simplifying the mortgage journey for homebuyers and lenders through intelligent automation and personalized guidance.**

Dwella is a mortgage automation platform designed to modernize the mortgage experience by reducing manual effort, simplifying complex processes, and connecting homebuyers with the right lenders. The platform leverages **AI and large language models (LLMs)** to guide users from discovery to pre-approval, creating a smoother experience for buyers while delivering better-qualified leads to mortgage professionals.

## Services

- LangChain
- LlamaIndex
- Pinecone
- LLM Integration
- Fine-tuned LLM
- AWS
- AI Agents
- Process Automation
- Flutter / Web

## Challenge

The mortgage process is notoriously fragmented, opaque, and time-consuming—particularly for first-time buyers. Lenders struggle with lead quality, document management, and client communication, while homebuyers face overwhelming decisions and a lack of transparency. Dwella aimed to solve both sides of this equation with an **AI-first platform** that automates workflows and delivers personalized support at scale.

## Approach

We built Dwella from the ground up, designing an intelligent workflow that automates loan discovery, borrower education, document collection, and lender-matching. Using custom-trained LLMs and real-time data integrations, the platform provides users with contextual guidance and next steps—while streamlining operations on the lender side with automated communication and qualification logic.

## Outcome

Launched an AI Mortgage Copilot prototype that guided home buyers through prequalification in minutes instead of days. Early pilots showed 50% faster preapprovals, a 35% increase in lender response rates, and improved borrower satisfaction scores by +20 points.

## What We Did

#### 01/Platform Strategy & AI Use Case Definition

Mapped out the full mortgage lifecycle and identified high-impact points for automation and AI integration—on both borrower and lender sides.

#### 02/LLM-Powered Borrower Assistant

Developed an AI agent trained to guide users through qualification, document prep, lender matching, and Q&A using contextual prompts and real-time updates.

#### 03/Lender Dashboard & Lead Routing System

Built a secure web dashboard for lenders to manage, communicate with, and track leads— prioritized by AI-driven scoring and readiness indicators.

#### 04/Automated Document Collection & Task Management

Integrated third-party tools for e-signature, ID verification, and financial data pulling, automated via workflows triggered by user behavior and AI analysis.