How reAlpha AI works

reAlpha AI diagram on how reAlpha works

reAlpha is on a mission to democratize real-estate investing by allowing fractional ownership of short-term rental properties through an easily accessible digital marketplace. One of the most important tasks in making this possible is to identify the best investment opportunities on the market. reAlpha relies on its proprietary algorithm, reAlpha BRAIN to source the best properties from the market. BRAIN analyzes a variety of factors for each property and assigns it a reAlpha Score, which is an indication of the viability of the property for the short-term rental market.

The reAlpha BRAIN comprises 6 major components that form the Artificial Intelligence powered system to identify the best investment opportunities in the short-term rental market. Lead Generation This is the component that sources the properties that are currently in the market for the system to analyze. The main sources it currently relies on are the emails sent by real-estate agents, properties listed for sale through reAlpha’s website and the data scraped from top real-estate listing websites.

Data Collection & Storage

This is the component that fetches the data for the identified properties from various third-party sources like data API providers, scraping real-estate listing websites, etc and aggregates the information into a central storage. reAlpha Score Property scoring is the core part of BRAIN’s infrastructure which is responsible for analyzing the property data and assigning it a reAlpha score. Humint This component is a platform used by human analysts to input data about a property using manual inspection and domain expertise, which was otherwise not possible for the data collection pipeline to fetch.

Admin Panel

This component is the web dashboard and control panel which the investment committee at reAlpha use to view the identified properties and make investment decisions. Humint Feature Predictor This component is the AI model that aims to replace the Humint component after sufficient training data has been collected for the model for it to be able to predict the property features requiring manual inspection with reasonable accuracy.


The typical journey of a property through the reAlpha BRAIN system looks like the following:

The property analysts use the Humint platform to provide data on the property features which require manual inspection (like exterior appeal, estimated rehab cost, estimated time to list it on Airbnb) Upon receiving the new information about the property from input, the Property Scoring System updates the reAlpha score.

The web dashboard in the admin shows the property data, photos, trend graphs and charts along with the reAlpha score for the investment committee. The decision of the investment committee to either buy or reject the property is used as feedback to update the Property Scoring System’s algorithm using Machine Learning. The property features input by the property analysts in Humint is used as training data to train the Humint Feature Predictor AI model which aims to replace the Humint after collecting a significant size of dataset to train the AI upto a reasonable accuracy.