UX Portfolio
cover.png

Prophia Abstract - AI Abstraction Tool

Prophia Abstract

Role: Sr. Product Designer (sole designer)
- I ran the whole design and product lifecycle from 0 to 1
Duration: 3 months
Team: 3 Engineers + 3 Data Scientists
Live Product Link: https://www.prophia.com/instant-lease-abstract

Overview
Prophia Abstract is a new B2C tool that lets users upload commercial leases and receive AI-generated summaries within minutes instead of hours and days. We saw an opportunity to bring our enterprise tools to smaller investors and empower them with faster, clearer lease data. Saving time and money!

Design Lifecycle (0 to 1)
- User research & design partner testing
- Wireframes, prototyping (v0, Lovable), and creating user flows
- Creating a new B2C-friendly design system
- Roadmapping, requirements, dev kickoffs + handoff

Research and Discovery

Full feature inventory

  • Registration & Login

    • Oauth (google, microsoft, okta)

    • Manual Account Creation and Fields

    • Verify password, reset password, password guidelines

  • Global Navigation

    • Side Navigation

    • Landing page

    • Account Settings

    • Support

  • Landing Page

    • Document Uploader

      • Single document upload and abstraction

      • Multi-document upload and abstraction

    • Table of Abstractions

      • Abstract State: In progress, Completed, Failed

  • Lease Abstract Summary

    • Columns and organizations of 40+ Terms

    • Customize each abstract - ability to select which terms to look for on documents

    • Single Document Abstract

      • Document Viewer

      • Page References

    • Multi-document abstract

      • Having Gemini track the order of the lease documents and pull the correct values for each term

        • Document change history tracked by AI

        • Individual term history

      • Document Viewer

    • Edit/Delete/Reset term values and references

    • Download/Export

  • Stripe Payments

  • Account Settings

    • Account Details

      • Change password

      • Delete Account

      • Add Oauth provider

    • Payment History

    • Promo Codes

Design System
Created a new design system that fits more of the B2C approach, to create components that are more friendly and approachable.

User Research

To do user research, we recruited prospective users and current users of our core product to be beta users and design partners. Due to being a team of one, I utilized tools such as v0 and Loveable to do easy prototyping to help share my idea to users and stakeholders.

  • Utilized moderated tests using prototypes and designs

  • Surveys - feedback and comments

  • Posthog - watching where users get stuck, how users are moving around the product, and any notable interactions

Example: https://v0-image-analysis-1lxn22.vercel.app/


Initial Iterations

I focused on simplicity and early alignment:

  • Built clear, intuitive flows before diving into UI

  • Used low-fidelity wireframes to get early buy-in from stakeholders

  • Iterated quickly based on feedback from real users

Notes:

  • To get some market research and reach, we started off with a free alpha version where we offered an excel export version of the lease abstract summary

  • We started with a single document abstract version soon after to get users in the door and to get user feedback early on the abstract summary

  • Worked hand in hand with the data science team to make sure we were fully utilizing all the data they were holding and processing

    • Overall document change history

    • Individual term history

    • Abstract term configuration

    • Document references for each current value and historical values


Current Version (LIVE)

Prophia Abstract launched just one month ago — and we’ve already seen exciting traction:

  • 50% user growth week over week since launch

  • We’re shipping based on real needs, informed by constant design partner feedback

  • Organic adoption is expanding as word spreads

We’re continuing to evolve the product, but the early signals are strong: this tool is solving real pain points, and doing it fast.