Glamour with a side of rain
When I decided to study abroad, I never thought my journey would start with a bang, and not the good kind. I was so excited about the university and program of my choice that housing was one of the last things on my mind. In my home country, finding a place to live matters, but it is rarely a struggle. Then I started looking, read about students who gave up their program because they could not find a room, and realized I was in for a long ride.
Over four months, I sent at least 300 applications, paid for five housing websites, called agents and almost got scammed twice. It was the battle of a lifetime, but in the end I had a place to live. Only in my second year, once I had friends to search with, did the chaos of the Dutch housing market become easier to navigate.
When I moved from my previous university city to Rotterdam for my master’s, I had no connections again. By then, technology had made its way into the housing market, and Stekkies brought all the listing websites together for me, which lowered both my costs and my stress. It was still a war, just on a smaller scale.
The team: four strangers, same household
Information Strategy was part of the first module of my Master’s at Erasmus, so our team for the AI Strategy Lab started as four strangers who had just met in class. Mateo and Joost live and work in Amsterdam, I live in Rotterdam, and Juline came from Paris as an exchange student. These different life experiences and situations made for a very interesting mix of perspectives while we brainstormed.
During those sessions, I could not help thinking how much easier my experience in the housing market would have been if I had not been alone. All those beautiful two- and three-bedroom apartments were nothing short of a dream for someone searching on their own. When I said this out loud, Mateo and Joost agreed, and Juline was just as interested, since finding a place for a stay of only six months is even harder. We all started working to turn it into a feasible, complete idea, and that is how Stekkies Together came up: an extension that keeps what Stekkies already does well, bringing the listings from all websites into one place, and adds matching with people who have similar lifestyles and budgets, so you do not have to face a new city alone.
My part: the demo, and what could go wrong
My most important task on the project was the demo. Together with Mateo, I worked out how the extension would work and designed its look and how people interact with it, from creating a profile and finding people to search with to forming a household through the matching algorithm. After the design was in place, we coded the functionalities using Claude Code, and then we also proceeded to include a market simulation.
The other half of my work was the governance and failure analysis: analysing where the extension could fail and working out how to keep it fair and legal. Because I have a technical background, I know how important limitations and governance are when building software. I think they are the foundation of the system: without them, your system is a photo without a frame. I mapped multiple failure modes and proposed safeguards for each and every one of them.
Key decisions and iterations
The most important decision we made along this journey came at the very beginning, when we chose to build Stekkies Together as an extension of Stekkies instead of building a new platform. We decided to target the part that was missing from Stekkies and all the other platforms: connecting possible renters with each other, which gives each user a bigger pool of homes. In line with Furr and Shipilov (2019), we did not try to disrupt the market, but to serve a need people already had in a better way.
Building the demo took several iterations. In my previous projects, I was encouraged to use AI to fix code, not to build it, so I was not fully used to this way of working. After the first prototype, I noticed small functionality issues: users could not leave a household or be part of several at once, and the matching stopped working properly after someone’s first match. This made me realise that the minute you leave a detail out, Claude tries to fill the gap itself, and there is a high chance it will hallucinate in the process. So I took a step back and wrote proper requirements with an extensive amount of detail, which made a difference in the following iterations.
In the end, we built a beautiful demo that my team and I are very proud of, one that properly brought to life the vision we had when we started this journey.
Demo Day: rock or rough diamond?
My team was part of the last session, after almost everyone had presented their ideas, and I must admit I was a bit anxious, both about pitching in front of a crowd and about the feedback that would follow. Of course, we believed in the quality of our product, but you never really know until someone outside your project bubble tries it out for themselves. Three groups came to our stand, listened to our pitch and tried our demo. From the very first moment, I could see the interest on their faces: they listened actively, asked many questions and pinpointed every little detail they wanted to try out, smiling and engaging with us. It made me very happy, and at the same time proud of the work my team had put in and the thought process behind the final product. On a side note, one of the students jokingly asked if we were going to pitch it to Stekkies themselves, to which I can only say: stay tuned for my future blogs.
This experience made me even more confident in our product and showed me that we had built something valuable that matches what our target audience wants. At the beginning of the session, our professor had presented the Philips Innovation Award, which I did not really pay attention to, but by the end of Demo Day, I had started to consider trying our luck for it.
How my understanding of AI and business strategy changed
I got used to AI and its coding capabilities early on, and in a way, I grew up with it. Still, building this demo surprised me. It may sound like a cliché, but this experience taught me that the sky is the limit: with the right idea and a proper plan, AI can bring most digital ideas to life, or at least help build them at a much faster pace. In a few weeks, we went from a brainstorm to a working demo.
However, as this course has taught me, AI is a widely available technology: everyone has access to it, so it rarely creates a lasting advantage on its own (Barney & Reeves, 2024). It is up to us to turn this new pace of development into value and leave our own mark. For Stekkies Together, the lasting advantage is not the AI algorithm itself, but the connections it helps facilitate.
Looking back at my first year in the Netherlands and my hundreds of applications, what I needed most was not to be alone. So, I am leaving this university project with the hope that Stekkies Together, or a technology with a similar aim, will actually appear on the market, whether or not I am one of its owners.
References
Barney, J. B., & Reeves, M. (2024). AI won’t give you a new sustainable advantage. Harvard Business Review, 102(5), 72–79. https://hbr.org/2024/09/ai-wont-give-you-a-new-sustainable-advantage
Furr, N., & Shipilov, A. (2019). Digital doesn’t have to be disruptive. Harvard Business Review, 97(4), 94–103. https://hbr.org/2019/07/digital-doesnt-have-to-be-disruptive
