Real Estate AI Platform
Background
Buying and selling a home without a real estate agent is a daunting and challenging task. There are many regulations and procedures involved to ensure the transaction goes successfully. It’s hard to keep track of everything to do.
Project Challenges
-
Limited timeframe and budget
The client needed to get an MVP built for a specific budget ASAP.
-
Limited skillset
Our team had little knowledge of working with AI or ML. They also had no training or experience working in a fast-paced prototyping environment. We also needed to incorporate 10 3rd party applications, some of which never worked properly.
-
Inexperienced founder
This was the founder’s first digital product venture and although he had many compelling ideas, did not know how to build or launch a product.
My role as Design Director
-
Develop product strategy
We had a small 5-person team and no dedicated product manager, project manager or account owner. I needed to serve those roles as well and as unite the team around a solid vision as soon as possible. I partnered with the founder to align on critical user journeys and a north star metric to create a tangible long-term vision.
-
Build and nurture a healthy team culture
The learning curves were steep, and the founder, eager to tackle numerous tasks simultaneously, lacked a technical background and was short on patience. Therefore, it was essential to establish a culture of transparency and trust from the start. This was my initial collaboration with the internal team, so I dedicated significant time to developing relationships with them and gaining insights into their challenges and superpowers.
-
Ensure design excellence
In order to meet the tight deadline, we leveraged the Ionic style guide and built on top of it. We created a robust, responsive framework that would allow us to quickly iterate on various concepts. I also initiated design reviews with the entire team to ensure a diverse perspective on our work.
-
Designing with AI
Prior to this project beginning, I spent months researching and test driving various AI tools to better understand how to optimize the design process. I published an internal doc that ignited a movement in the company to leverage AI tooling in key workstreams.
-
Create a robust research program
Led an onsite workshop and research days to better understand the nuances of the workflow of high-performing real estate agents. We identified gaps in existing tools, fragmented user experiences, and numerous workarounds to expedite processes and optimize time with actual clients.
-
Design for AI
Designing for products that leverage AI required a deeper understanding of how generative AI and other features operated. I leveraged existing design methods and created a new persona based model to ensure we understood how to create the ideal experience when we had so many unknowns.
Our Process
I led the effort by establishing a detailed desk research program. We leveraged AI tools like Chat GPT and Perplexity to understand domain terminology, pain points, competitive landscape, and workflows.
Understand the domain
We used tools like Dovetail to automatically transcribe our recordings and analyze recurring themes.
I spearheaded our first AI-driven research initiative, employing AI tools to design interview sets and organize interviews with buyers and sellers to explore the details of their daily challenges. Additionally, I developed a custom chatbot based on our research findings to enable rapid analysis and broaden the reach of our knowledge.
AI based analysis
I led a design workshop to identify and ideate opportunities across a complex user map.
We also developed a north star metric of the number of listings so we decided to focus on differentiating and optimizing that experience.
Design workshop
We worked closely with the client to better understand what the brand means to him and how we can create a distinct look, but still let the content shine.
Explore visual directions
Define the right experience
As we explored visual directions, I also led an effort to create the ultimate experience for real estate agents. We wanted to maximize their time with clients and minimize their time with technology.
We created interactive prototypes and ran remote usability studies to validate decisions.
Crafting a vision
Once we were grounded in the domain, we created a vision for what this product could be in the near future. We leveraged new technologies and optimized for simplicy and focus.
Making it real
I collaborated with several developers new to AI technologies and rapid prototyping to develop a key feature of our demo called "instant listings." Utilizing address data and property photos, we crafted a comprehensive listing that needed minimal adjustments.
The challenge of building this prototype quickly with a novice team required daily coaching and setting clear weekly goals. I facilitated numerous discussions to address obstacles and swiftly negotiated technical compromises. Moreover, I aimed to support and encourage the team, fostering confidence while managing stress levels effectively.
The Impact
We launched in the Google Play, Apple app store and on web at the end of January 2025. Our client has extended us twice and we continue to evolve the app and work with marketing to build a strong user base.
The time to create a listing has been improved by 50% and the buyer experience has been greatly simplifed with proactive notifications of property matches and transaction support for a fraction of the cost of a real estate agent.
What I learned.
It was my first time working with this development team and I made some assumptions about their skills and comfort with ambiguity. In the future, it’s critical to align approaches and mindsets before starting rapid prototyping work since it’s dramatically different from building an MVP.
Timeboxing is even more critical in projects with limited timeframes and resources. Daily check-ins could have begun earlier as it took a while for developers to clearly articulate the impact of their tech platform choices.