Why we built zelios
Realtors were losing hours to something a computer should do
After talking to many realtors in our area, everyone had curiously developed the same process to classify their real estate pictures. They would manually change the names of the files on their computer with numbers to order them in the way they wanted before uploading them to their real estate platform of choice.
Finding a way to classify pictures easily
Naturally, AI was the first thing we tried. However, we were aware that giving too much freedom to any AI model is a sure way of getting unreliable results.
This is why we developed a process which does not fully take away humans. A process where the AI has exactly one job which is: "Hey, tell me about this image." Then realtors can validate if the AI has made mistakes through our validation step.
A bit about who's building this
I am a computer engineering student who happens to love building stuff. Video games, music, websites, etc.
The main reason I decided to pursue the real estate image classification path is because I took the time to have conversations with many realtors who all passionately told me how much they hated this part of their job.
For some realtors, we're talking about 2-3 hours of manual classification every week! It sounded so crazy to me that I had to go full throttle on this project.
Where this is headed
We're working on making classification faster, more customizable, and even more private over time.
