Get our exclusive interviews & blogs in your inbox!
Photography

Listen to your gut, develop a plan, and go with it - Story of Foxy AI

Startup Name
Foxy AI

Revenue
$ 1000000 MRR

Location
United States

No Of Founders
1

No Of Employees
3

Hello ! Who are you and what are you working on?

I am a former real estate developer turned tech entrepreneur. Foxy AI is working to empower valuation modelers, financial institutions, and app developers with visual property intelligence by incorporating data from your property photos.

What motivated you to get started with? How did you come up with the idea?

I had been following the AI tech scene for quite some time before starting Foxy AI. I was looking for an opportunity to apply artificial intelligence to real estate in a way that could provide immediate value to a customer. We had an appraisal on a new development that came in below where we expected it would be. Fortunately at the end of the day it all worked out, but it got me thinking about other ways of valuing real estate. That's when I came across Automated Valuation Models (AVM). The most famous of which is Zillow's Zestimate. After researching AVMs I learned that the one of the major draw backs to their use is they cannot account for the quality and condition of a property. I had an "Aha!" moment. Turning property photos into usable data containing information on the quality and condition of a property was the AI use case I was looking for.

Can you tell us the story of your business from idea to where you are now?

I began to research computer vision in dept to gain a greater understanding of the technology. I was introduced to my partner and technical adviser, Frankie Liuzzi, who built our first neural network which we call house2vec. House2vec takes residential property photos and turns them into an image feature vector, an array of numbers representing features within the images relating to the quality and condition of the property. This data can then be incorporated into an Automated Valuation Model to improve accuracy in much the same way that other features, like the number of bedrooms and bathrooms, are used. Since developing house2vec we have added a room classification network which has become an important part of our neural pipeline. 

Most recently, thanks to feedback from our partners, we have built an end to end condition scoring application which has quickly become our most popular tool. 

Some of the companies we are working with said, “Hey, we are missing a big piece of information, we don't have a way to determine the condition of the property in an objective, standardized, and automated way”. Of course the condition of the property is such an important piece of the equation and the only way to get reliable information is to either physically visit the property, which isn’t always possible, or look at photos of the property, which is time consuming at scale, and both are subjective. 

So we took took our valuation technology and put together an end to end solution to score the condition of the property, based on the photos, with an easy to use, easy to understand score. We based this score on the scoring system from the Uniform Appraisal Dataset which Fannie Mae, Freddie Mac and others use for underwriting. So now, using artificial intelligence, we have the ability to determine the condition of the property automatically, at scale, using property photos.

What has been your biggest failure or struggle?

Our biggest struggle has been educating business people on the value derived from use cases specific to their industry. We are building a new class of products, something that didn't exist before - visual property intelligence. This requires us to evangelize the problems we are solving, and then provide them with the solutions. 

And what has been your biggest achievement or success?

Recognizing and addressing this issue of abstractness around our house2vec product and as a result, building the Condition Score as an end to end solution has been one of our biggest internal successes. The Condition Scoring tool has opened us up to a much wider variety of companies. From investors to listing aggregators, data aggregators, appraisers and everyone in between, they are all finding interesting and unique use cases for the Condition Score in their businesses. 

Since launch, what has worked to attract and retain customers?

The basics of any business are most effective. Delivering on your promises, remaining flexible, being prompt, and providing the best service possible are what keeps customers on board. They know we genuinely care about helping them improve their business while building ours. 

How your real estate background help you in the development and selling of Foxy AI?

Real estate development, and a natural interest in science and technology, led me directly to starting Foxy AI. When an appraisal came in lower than expected I knew there had to be a way to improve the valuation process. 

Many of our customers are real estate investors like myself. Understanding their business process and pain points are key elements when building and selling products. 

Did you use Betalist or PH or other Startup Launching Platform for Launching ? How was that experience ?

No we didnt use any platforms when launching. As a b2b business in a specific niche, many of the users of these sites dont apply. 

What’s your business model, and how have you grown your revenue?

Customers are charged on a per API call basis at $1/API call for our Condition Score tool. You can contact us for a full list of API features and pricing. 

Our goal is to make our tools available for use to companies of all sizes. Foxy AI will be the backend provider of visual property intelligence tools to the real estate world. One of our main drivers of revenue is working with our customers to explore non-obvious ways of utilizing our tools and data within their business. This helps them increase efficiencies and generate more revenue as well as helps us grow our business with existing customers. We are heavily focused on relationship building for the long term. 

What are the biggest challenges you’ve faced and obstacles you’ve overcome? What are your goals for the future?

Foxy AI is focused on visual property intelligence for residential estate. As we move forward with the business we will continue to expand our tools for visual property intelligence. We are collaborating with our partners to add new features to our current offerings, build larger datasets, and improve the accuracy of our current models. We have some exciting new tools in the works!

We are very excited for our Condition Score application. As an end to end solution, this has expanded our customer base from AVM providers to many other areas like iBuyers, investors, insurance companies and marketers. 



Our goal is to make our tools available to everyone which allows our customers to explore new and creative uses within their business.  

How are you doing today and what does the future look like?

The future is a world where we have moved beyond the basics of property information. Knowing just the size, location and number of bedrooms just isn’t going to cut it. Consumers and investors want to know what types of rooms make up the property, the condition of the property, what features are most important to consumers in this particular town, etc., and they want to know immediately. 

Through starting the business, have you learned anything particularly helpful or advantageous?

Like real estate development, software development takes longer than you expect.

What have been the most influential books, podcasts, or other resources?

Podcast: Peter Diamandis's Exponential Wisdom

The best book I have read recently is David Goggins "Cant Hurt Me". This book will kick your ass into high gear.

What’s your advice for fellow aspiring entrepreneur who are just starting out?

Listen to your gut, develop a plan, and go with it. It will never be perfect and not everyone will like it. There is no failure, only learning.

Get our exclusive interviews & blogs in your inbox!