Year : 2024 Role : Lead Product Designer Industry : PropTech / AI / LLM Agents Team : Startup Timeline : 1 year
Impact at a glance
The Start-up challenge
At Nophin, our mission was to explore how AI could revolutionize commercial real estate workflows. Our focus was to assess the viability of our product for deal screening and underwriting, determining its potential to streamline traditionally time-consuming processes in CRE.
Results
1. 120% MoM growth in active users 2. Secured funding from top VCs, driving toward $100K MRR 3. Transformed a SaaS platform into an AI-powered CRE analyst
Exploring the goals
Unlocking AI’s Potential in CRE
Commercial real estate (CRE) deal screening is slow, manual, and data-heavy. Acquisition teams spend hours analyzing rent rolls, T-12s, and market reports. Our goals included :
Make research faster
Automates financial analysis & market research
Automate underwriting workflows
Accelerates deal screening & underwriting
Provide data insights
Enhances decision-making with real-time insights
From Concept to Reality
Designing Cresa, the AI CRE Analyst
We rapidly iterated to build Cresa, an LLM-powered deal screening tool that:
1. Extracts key financial data from CRE documents 2. Syncs with real-time market data for instant benchmarking 3. Automates scenario analyses & memo generation
Result: Teams screened deals 10x faster, focusing on high-value opportunities.
Scaling Innovation: UX-Driven Growth
User-first AI workflows that felt intuitive and powerful
Tight collaboration with engineers & analysts for seamless integration
Continuous iteration based on user feedback & ML performance
Results
Innovation was key to our success
By combining UX with AI, we redefined how acquisition teams approach deal screening. The success of NophinAI proves that great design is the bridge between AI innovation and real-world impact.
Expand our team & user base
Creative technology enhances interactivity and user engagement.
Raise multiple funding rounds
I prioritize data analysis to guide design choices.
Establish Cresa as an industry-first solution
I ensure a smooth transition through every phase.
Design + Machine Learning
Optimizing the AI Model
Building a powerful AI analyst wasn't just about automation—it was about ensuring accuracy, usability, and trust in Cresa’s decision-making. I worked closely with ML engineers to refine the model using a UX-driven approach that turned raw AI outputs into actionable, human-centered insights.
Working closely with the ML team
Data-Informed UX for Better AI Performance
Refining Input Models
Designed structured data inputs and user flows to minimize ambiguity in AI interpretations, ensuring higher accuracy in financial extractions from rent rolls, T-12s, and underwriting documents.
Enhancing Feedback Loops
Integrated real-time user feedback into model outputs, allowing acquisition teams to correct AI errors on the fly, improving model confidence and precision over time.
Optimizing Usability
Simplified complex AI outputs into digestible, actionable insights, balancing automation with transparency so CRE professionals could trust and easily verify AI-driven recommendations.
Personalization
Developed an adaptive system where users could customize screening criteria, weight financial metrics, and refine AI assumptions to align with firm-specific investment strategies.
Conclusion
By embedding human oversight within AI workflows, we increased model adoption, improved decision accuracy, and gave users a sense of control—critical for AI adoption in a high-stakes industry like CRE.
120%
MoM growth in active users
$100K+
MRR milestone secured through AI-driven efficiency
10x faster
deal screening with AI-powered automation
80%
reduction in manual data entry for acquisition teams
3x increase
in underwriting speed with instant financial extractions
35%
improved AI accuracy through UX-driven model refinement
