The structure of your report will be similar to what is used in industry and will follow this outline:
Project Goals
I’ve been in real estate for almost 2 years now and I run a remote property management and outreach company that helps real estate investors. As I met more people and worked with new clients I discovered a theme, real estate investors don’t have a modern way of tracking their numbers. Most investors use spreadsheets to track and do their calculations which leaves lots of room for error. When it comes to an assest like real estate, there is never a $1,000 mistake it’s always 10,000 and 100,000 in mistakes. I decided to create a website that modernizes real estate portfolio tracking and analysis. Later in the project, due to a small target market I decided to expand the website to include a calculator. Allowing the website to be useful towards all real estate investors.
Project Constraints
My users include all sorts of real estate investors. The calculator assists new investors with purchasing and analyzing potential properties while the portfolio manager helps more seasoned investors with tracking their income and debt. In fusion I only had around 2-3 months to complete the project, but I will be extending that timeline into next year because I aim to run the website as an actual business. The main constraint when I first approached this project was skill. I didn’t know how to use figma, bubble, or any of the other website creation/design websites. Learning to use figma took out a significant amount of time and therefore I could only produce a design concept for the website. However, now that I’m progessing with the website I’ve found a group of developers at UVic to properly develop it.
Project Expectations
Ideally, the website is sustainable and can run as a fully feldged business. Since there is still a while before the website can be fully developed I would base my definition of success on the amount of sign ups that we get and the amount of angel investors that participate in the project.
Design Choices
The solution as I mentioned above is a website that encompasses the entire journey of a real estate investor. Starting with a free tier calculator and moving onto a paid tier portfolio manager. What I didn’t mention was an AI add on as a premium tier. There are currently already lots of different calculators and property management softwares but they don’t encompass all the different creative strategies that investors nomrally use. This is why I’m incorporating an assistive chatbot that will help these investors with coming up with strategies.
These are the concept designs I had for the portfolio manager. Designed in figma.
We are most likely going to change up the AI design as the chatbot wouldn’t function as an AI but more like a analyzer. It will be coded to only know answers to questions related to real estate.
Project Planning
The first step was identifying wheter there was a demand for a website like this. I reached out to my clients and to my friends and was then confident enough to continue with the idea.
Next I had to choose a way of creating a prototype. I found that I could either use figma or bubble to create a basic website. However, these websites were not easy to use and I don’t have the necessary skill which resulted in me having to downsize the project and create a more simple MVP.
Now with a more simple MVP in mind I took the time to learn how to use figma and created the design I showed above. But I still completed ahead of time so decided to take some time and do some reasearch regarding the AI I wanted to put in the website.
I didn’t know how to code or develop an AI nor did I have the time to do so, so instead, I did some reaserch using the Open API platform. I will show a screenshot of the code I used to “prompt” the AI but basically used the model to test out some question and answers. This is the link to the full test. To give a breif summary, AI is extremely good at finding possible strategies and doing fast caluclations. However, the main problem is that AI is inclined to give advice even when it may not be qualified to give such as legal and financial advice. The solution to that was to put hard limits on the AI to restrict it from engaging in any conversation about legalities.
As I mentioned, there are still more steps to the project but I will mention it later as this is the full extent of what I was able to complete in Fusion 10.
After creating a slide deck with the design, I started asking people around in my network what they thought of it. Many liked the simple and organized idea but felt that the target market was small. Many said, “How many people do you know own a porfolio size that big?”. After giving it some thought, I decided to expand the website to include a free tier calculator that beginners could use. I also worked with the development team to find solutions to a working investor chatbot to limit liability concerns. I’ve explained it more indepth below.
Bill of Materials and Cost Analysis
Figma was free which meant the only thing that costed some money was purchasing tickets to prompt the AI. I’ve put in $20 but still have over $19 left over. After all, I didn’t really use the AI during any of the presentations due to the lack of organization.
Moving on, maintaining and developing the website will take a lot more of a financial strain then the MVP. So far the team gave me a development cost of $10,000 and a variety of maitinence packages ranging from $500 to $5,000. I plan on sourcing angel investors for the initial development and then pitching it to more investors after the website is “investable”.
Testing and Validation
I belive that the solution will be successful because many of my investors have expressed an interest in a website like this. I’m also in the process of getting presignups and have over 20 right now. The MVP isn’t exactly testable because it’s only a design concept but I’ve been pitching the project to multiple investors. If they’re willing to come in as an angel investor, I would consider that I major success.
Code of Ethics
The main problem regarding Ethics about the website is the use of AI that relates to any sort of financial gain or purchase. First of all, AI does not have the qualifications to advise any financial transaction as well as any legalities. Purchasing real estate involves a lot of contracts and also involves lawyers, realtors, and other professionals. Should our AI respond in anyway that engages any of those conversations, it would put the company at a huge liability risk.
Intially, I had thought that we could simply limit the AI to certain responses but later discovered that the user then wouldn’t truly be able to maximize the feature this way.
The developers suggested their own plan. They said, “In short, we will reduce the AI’s hallucination by providing it with a limited and heavily bounded knowledge base. Hallucinations occur when the models has too much knowledge and hence, may stray off topic leading to inconsistent and incorrect responses.” Our knowledge will be manually written in documents which will be exported as PDFs. Those PDFs will be converted to vectors which will allow the model to search for information (similar to how google searches for articles when you search for keywords). Once the models pulls what it deems to be the most relevant information, the model will generate a fluent response for the user. So in conclusion, the hallucinations that may cause liabilities can be used by limiting the knowledge base of the model and only using the AI’s conversational abilities.”
To a teenager without lots of experience I believe they would know how to handle that probelm a lot better than me.
Conclusions and Recommendations:
As we move on with the project we are trying to gather more funds to launch the website periodically and create the demand for our software. Here’s a basic timeline of the next year.
Pre Summer Break: Initial landing page + pitch presentation assembled
Sept 2025: 1st round fundraising ends, ideally, enough to start our first development
Sept to Nov: First development stage
Nov to Jan 2026: Advertising + getting clients, seeing if people like the website
Jan to Feb: Fixes/improvements + 2nd round fundraising
Feb to May: Second development stage (hopefully all the development is done)
May to July: Official launch of all features -> extensive marketing -> attaining clients
July to Sept: Last improvements/feedback -> Purchase some of the more advanced maintience packages, etc.
Luckily, we seemed to have found an investor willing to put in the full $10,000 to fund the development. Along with the signups, the project seems to have bright future ahead.
References:
None. Everything was a recall of what happened this past couple of months as well as my insight on the project except for that one excerpt from a developer. Special thanks to all the investors that assisted me with the project as well as the developers that gave me the opportunity to explore this project even further.
Appendices
This is the code used to prompt the AI.
Where is says “input” is where I would plug in the tests.
I’ve been using Motion to keep myself on time and organized throughout the entire project. There is a detailed explanation in this blog post. Here are some snippits. (motion please sponsor me)
Thanks for following along on my journey and if you would like to get into real estate feel free to call me at (888) 240-3598. Instanttenant signing off.
Leave a Reply