How this guy used AI to Acquire 15 Off-Market Short-Term Rentals in Just 12 Months

[Mark Simpson]

After reflecting on a strategy someone used to buy 15 short-term rentals off-market in just 12 months, I thought of an idea. Here’s how they approached it:

1. Scraping Listings: They used web scraping tools like Beautiful Soup and GPT APIs to scan Airbnb for properties in their target areas.

2. Image Analysis: They trained an AI model to analyse listing photos, identifying homes with solid potential (good structure, layout, location) that were undervalued due to low-quality photos. These properties often generated decent revenue but underperformed simply because they didn’t stand out visually.

3. Natural Language Processing: They also trained an NLP model to review poorly written Airbnb descriptions. The model flagged listings that could likely generate more revenue with better descriptions.

4. Market Comparison: Using Zillow Zestimate APIs and AirDNA, they compared similar properties in the area and forecasted how much more the property could make if the photos and descriptions were improved.

5. Property List: All of this data was organised into a Google Sheet for further review.

6. Automating Outreach: They used Claude to generate personalised outreach emails and DMs to homeowners, pitching offers for the properties.

The entire process revolves around identifying undervalued homes that look worse online than they actually are.

Has anyone seen anything like this?

[Andrew Steffens]

Interesting but complex.  In the same amount of work I could find an undervalued deal in my market through existing channels.