These two ladies are developing Oreeva and using artificial intelligence to make property evaluation smarter and faster
Just last month I had a conversation with Lidia and Roksana from Oreeva.com over skype to get a better grasp of the project they are working on. With artificial intelligence, and its potential impacts in people’s lives being subject to much debate these days, nothing like understanding a bit more how this technology is being used to substantially alter time consuming processes such as getting a price tag for a piece of real estate.
This project and its current state wasn’t born overnight. In fact it was the result of a somewhat serendipitous meeting between its two founders, Lidia and Roksana, and subsequent iterations on their initial ideas — some of them even taking place during the period in which this article was being written — to get to its current configuration.
A meeting of minds
Lidia and Roksana met in Zurich some years ago and found there was one thing they had in common. Lidia had worked in finance for the most part of her career and developed an interest for artificial intelligence. Roksana, also from the finance industry, was devoting some of her spare time to buying, improving and selling properties and was interested in a tool that could be used to make this process easier and more transparent. Uniting the love for the technology and the desire to use it to solve a real world problem, the conditions were met for the two to start brainstorming about a new product.
Even this interest in the new technology wasn’t born out of thin air. While still having a day job at Citi, Lidia used some of her weekends to help out a team of two Phd students that were developing a tool for asset management using deep learning.
“I gave them my insights as an industry practitioner and helped out with business development, while in return I got exposed to deep learning, something I was interested in.”
When the two founders to-be sat down to discuss their ideas, they came to the concept that they wanted to create nothing short of a “Bloomberg for real estate evaluations” that would allow all real estate market players access to the same accurate property value estimates, as well as property insights and predictive analytics, all at their fingertips.
The problem
Their starting point was that today many real estate evaluations are still being made using traditional statistical models, with an error margin that, it is being proven, can be decreased using models that leverage deep learning. In fact, it is still a reality having someone go on premise to make the estimates, which gives a notion of the sheer number of different valuations that can come up associated to each site.
While some countries are more sophisticated than others (like the US, Sweden, or Norway), there are countries, like the UK, in which, they have observed, these more advanced models are still not the norm.
In these countries, they claim that solutions like this, that apply deep learning to real estate evaluations and analytics could help to increase the accuracy of the estimates as well as provide indication as to its future value, an additional feature to some current models.
The MVPs
After agreeing on the concept the next step was to develop the algorithm that would be at the core of the solution provided. Through their personal network they found Artem — a New York based expert in robotics and deep learning that spent 8 months working full time on the algorithm together with Lidia and another researcher. He is now continuing to support the team in advisory role.
The team was further supported by a developer to build the algorithm, a software engineer and a team of front-end developers, that together with Roksana worked over a few months to develop this first version of the product.
The interest for this concept was first tested in the Singapore market using google ads and observing the click rate and inviting some users to experience the platform, whose feedback was then used to make product improvements (e.g. customers made observations about improvements to be made on the graphics).
Nevertheless, with one of the founders being based in the UK and another in Switzerland, the Singapore launch was temporarily deprioritized, and a launch in the UK under is now in the works, in which the focus is less B2C — individuals looking to buy or invest in properties and want to know estimates of their value — and more B2B — lenders or mortgage professionals that need these estimates for lending — contrary to what been expected and planned for the first market as this turned out to seem a much stronger pain point in the UK.
By now the team has grown and the technical part is led by CTO who has joined as a co-founder. They have also added a Berlin-based researcher with several years of experience in deep learning, who is now an advisor to the team.
Prospecting
The founders identified three broad segments for this product: lenders and brokers, real estate agents and valuers, and real estate investors. So these have been the players they have been trying to get in touch with.
Digital banks and online mortgage brokers, due to their core affinity with technology, seem more receptive to the introduction more advanced valuation models. For the first, the key people to reach out to would be the Head of Mortgages or the Head of Innovation, although they claim that sometimes some of the functions are somewhat separated from business-as-usual to be relevant for this product. In the case of online businesses, they reach out directly to the CEO or business owner.
Obstacles
During our conversation, and acknowledging the different levels of sophistication of the usage of the technology in different countries, namely in Northern Europe, that seems a bit ahead, there was a question about why the players that have developed those technologies haven’t entered countries such as the UK.
While this still remains to be fully understood, it seems true that “ companies using the legacy valuation models — that have invested heavily in implementing them and embedding them into their processes — are not so easily willing to replace them for a new technology, even under the promise of more accurate estimates.”
Communication and distribution
The communication so far has been made mainly by participating in conferences such as Proptech UK back in May this year and social media, but there will be more intensive efforts upon the launch in the UK market that is expected to be soon.
Future revenue streams
Oreeva plans to charge a price per valuation being requested on the platform below what is currently being charged by competitors, a fee per pilot project and a monthly subscription fee for access to data respecting a specific UK region.
Project status
After testing the concept in the Singapore market, Lidia and Roksana are now starting beta testing in the UK market. They made a conscious decision to develop the product together with customers and are fortunate to already have a few prospect clients who are happy and willing to give user feedback.
Since we had the first conversation they completed a major revision of their website and brand identity, whose latest version is already available to the public. The new brand, Oreeva, stands for online real estate estimates, valuation and analysis.
Initially self funding the project, they are now in discussions with a few strategic investors and VCs, to help scale the business, having also presented their idea at Inseaders Demo Day, on the 9th September.
It was quite interesting to hear about how artificial intelligence is being leveraged to make the everyone’s lives easier. Looking forward to hear about their progress sometime soon!
Reach out
You can check out their latest website at www.oreeva.com
If you are interested in understanding more specifics about their company reach out directly to them at lidia@oreeva.com and roksana@oreeva.com, or in case you already a member of the Inseaders platform you can login and view their deck at the Investors Lounge.
Oreeva was previously know as Providens.ai.
