As it often happens with entrepreneurs – a special breed of problem-solvers with enough drive to turn insights into actions – Ranime El Skaff decided in 2022 to create Prop-AI to address pain points that she and co-founder Christian Kunz had personally faced: the mass of low-quality, incomplete, and poorly structured information that anyone looking to invest in real estate has to sort through. The artificial intelligence-powered platform, officially launched last December, facilitates the search and buying process down to the minute details, from negotiation to paperwork.
What drove you to launch Prop-AI?
I had been wanting to invest in real estate for quite some time, but I barely had any time and didn’t know where and how to start. Being super analytical, I was building a spreadsheet but the deeper I was going, the more I was realising there was a lot more to explore. You can access a lot of information if you dig deep, but this information is not clean or clear. It doesn’t always answer your questions.
It turned out that my co-founder was having the same problem; so, we started thinking about developing a solution to help busy professionals identify the best real estate opportunity for investments.
How does the product work?
You just enter your preferences on our website and our algorithm works in the background, scanning the entire market. We analyse almost 500,000 listings a week, clean them up, score them, rank them, and make suggestions to the investor.
We use AI throughout the journey to analyse the data and get insights out of it – for example, whether a property is properly priced or not. We look into almost 30 different metrics at a high level and within those, we look into even more metrics, all the way from area, development, name of the building, floor, size of the building, and number of bedrooms, to current mortgage rates, currency appreciation, inflation, taxes, maintenance, etc. – things that a normal mind is unable to all account for at the same time.
This is why it irks me a little bit when people just use the property price or the rental yield to calculate return on investment (ROI) on real estate investment because there are so many factors to consider.
Which AI model are you using and how do you train it?
We use OpenAI for data cleaning, data processing, insights generation, etc. We’re also looking into Gen AI as part of our use cases going forward, to develop real estate agent profiles to source properties, negotiate, get the information that is needed, and so on.
To understand the fair market value, for example, we take all of the Dubai Land Department data and we enrich it with our own data sets – we have around 1.2 million data points that we get from aggregators’ listings. We check, clean up, and link everything to the end data so that they speak the same language. Then, we feed all of it into the model. We train the model on 80%, and we test on 20%. The first time we tested it, we had 40% accuracy. We made calibrations and we’re now at 92% accuracy.
What was the entrepreneurial journey like for you, leading a highly tech-driven business?
I have a Bachelor's in biology and a Master's in Health Management and Policy, which is very different from what I’m doing there. I didn’t know what I wanted to do so I went into consulting at McKinsey and then to a private equity fund until I decided to do my own thing.
I always had this itch to start my own company but I didn’t really understand how difficult it was going to be. My previous job was highly stressful but ownership is a different type of stress. Someone who had started two businesses asked me 18 months ago, “How are you feeling?” I said, “I get overwhelmed every now and then.” He asked how often and I said once a week. He replied, “Call me when it’s once a day.” I thought he was exaggerating. He was not.
I’ve been through multiple phases and each had a new challenge. In the beginning, when it was just me full-time, it was the discipline, the planning, the organisation, the ability to push and to find the balance between the dream and the vision. And it felt a bit lonely, but I’ve been lucky to have people helping me.
Then, you have a team but it’s your baby and you don’t know how much you can delegate to other people. That changed when I realised that the team was as motivated as me, if not more.
Later on, it was making sure the team gets paid at the end of the month, which is a responsibility I’d never had before. We raised half a million dollars in pre-seed from friends and family, and we’re lucky to have angel investors who believe in us, but the fundraising situation is a bit difficult now, with banks giving 6% interest rates. This is why we’re going to raise our next round only after we start generating revenues in a month or so – we have a very solid and healthy funnel at the moment. One important learning for me is that you never stop fundraising; it’s a continuous process.
Another big challenge is being able to wear multiple hats because you get trained for so long on doing one thing. In all fairness, 90% of the work is things that you don’t know. But it’s a learning curve and it’s also super energising; I’m very happy with the decision I made.
What other plans do you have in the pipeline?
We’re working on a forecasting model – for example, the impact of building a big landmark like a Dubai Mall in another area, of a recession, of an increase in mortgage rates, or of the growth of population, etc.
Also, for now, our algorithm only covers residential but we’ll go more into commercial at a later stage.
In addition, our active revenue model is commission-based, but we’re looking into other revenue models like subscription down the line, at the B2C level for portfolio management and the B2C level for data monetisation.
And finally, we have started building partnerships to expand to Portugal and Spain. However, Saudi right now is also very exciting, so we’re figuring out whether the next steps for us would be Europe or Saudi Arabia.
How do you see technologies like yours disrupting the real estate industry?
When a market is that big and doing so well, it’s very hard to change. That’s why real estate hasn’t been disrupted significantly in the past ten years. But the housing boom has ended and, with higher mortgage rates, different segments of customers are emerging – more analytical, not making rushed decisions.
It takes around 300 hours of research using the existing information, which a lot of people don’t have the time for. Or you have to trust an agent, who’s not incentivised by what’s good for you but more by what’s good for them.
We want to shift that power dynamic from the seller and the broker to the actual buyer. The information asymmetry is going to disappear, and people will feel more confident making investments that make sense to them.