What the AI Strategy Lab Taught Me
Before starting the team project, we agreed that every member would pitch one or two ideas to the group, so we could discuss them and choose our preferred one. While I proposed the idea of integrating agentic AI into the existing services of Funda, one of my teammates suggested developing a new AI-powered home-search agent that could search for properties across various platforms. It became clear that both ideas addressed the same underlying problem: the scarcity of housing in the Dutch housing market.
We realised that all four members of our team have experienced difficulties finding housing to some extent. In my case, moving from Groningen to Rotterdam to pursue my master’s degree came with several challenges. There were only a few student rooms available, so I had to constantly check multiple listing websites to make sure I did not miss any new listings. As I was specifically looking for a student room in a quiet street, within walking distance of a supermarket, and with an easy commute to the university campus, I had to manually check whether a new listing met my requirements.
This is where we saw an opportunity: we decided to combine both ideas by incorporating the strongest elements of each. From my idea, we took the ability of agentic AI to account for soft preferences, such as an easy commute to campus or living close to a supermarket. From the other idea, we took the concept of scraping listing websites to bring all available properties together in one place. This led us to the decision to develop a comprehensive platform that acts as a virtual real estate agent, which we decided to call “HomeMatch AI”. Homeseekers describe their housing needs in detail, and HomeMatch returns all accessible properties across various portals.
In the team project, it was my responsibility to clearly describe the problem as well as the opportunity behind it. It made me realise that the national housing shortage in the Netherlands does not only create challenges in the rental market, but also in the owner-occupied market. In both markets, the limited housing supply drives up prices and leads to increased competition among homeseekers, making it increasingly difficult to find suitable housing in the Netherlands (Pararius, 2026; Sitanala, 2026). I also noticed that property listings are spread across numerous listing portals and real estate agency websites, causing the search process to be even more time-consuming. It is clear that our project attempts to tackle a real and pressing problem in the Netherlands.
Alongside our decision to merge the two ideas, another important iteration was to shift our focus from students, the group we initially wanted to serve, to young professionals and expats. At first, focussing on students seemed like the most logical choice, as we, as a team, all know many students in our immediate surroundings who struggle to find suitable accommodation. However, when we discussed this topic in a later stage, we decided to redirect our attention to professionals and expats. This is because this group is not as familiar with the existing housing platforms in the Netherlands and will most likely have a higher willingness to pay for our digital service compared to students.
As I was the first to start working on the group project, it was my responsibility to choose the design of the slide deck, create the cover page, describe the problem and the opportunity, but also describe the competitive landscape our startup would find itself in. Through my recent experience with finding housing in the Netherlands, I was already familiar with most of the potential competitors. Funda and Pararius, for example, are known as two of the largest property listing platforms in the Netherlands. Alert tools, such as Stekkies, Rentbird and RentSlam, could be identified as more direct competitors, as they are situated in a similar spot in the ecosystem as HomeMatch AI (Renthaven, 2026).
From one perspective, the assignment showed me the importance of a well-developed business plan. While researching for the competitive analysis, the ecosystem map, and the future-market test, I realised how vulnerable the position of HomeMatch is in a scenario where our competitors start using agentic AI in their operations as well. This proved to be the most significant challenge we encountered. We decided to expand our idea by integrating practical real estate agent advice on how to proceed after a rent or purchase agreement is signed. It became imperative that HomeMatch AI would have to evolve into the best value-for-money home-search platform on the market. By learning from successful and unsuccessful search results, the model must be trained to truly understand the home seekers’ needs, find the right matches, and gain trust from subscribers. I realised that building a business involves a lot more than just developing a product or service, it also requires continuous development, competitive pricing, and establishing a trustworthy brand that customers can rely on. The project allowed me to experience this firsthand.
From another perspective, the project opened my eyes to the possibilities of artificial intelligence and what can be achieved with the technology available today. Initially, I was hesitant and questioned whether our idea was feasible and whether developing a prototype would be too challenging for us. However, I realised that, with the right tools, it is possible to turn a business idea into a working prototype. This experience changed my perspective on what can be achieved with AI, and showed me how accessible the development of technology-driven solutions has become. Especially when I saw the prototypes of the other groups on the Demo Day, I came to understand that you can create a lot more with publicly available AI tools than I had expected.
However, it is important to note that our prototype could not be coded by just inserting a single prompt into an artificial intelligence model. Many iterations had to be made in order to improve and eventually fine-tune the prototype. These changes can all be found in the appendix of the slide deck. I learned that AI can be used a tool to build a digital business, but that it remains essential for us, as humans, to keep improving and overseeing the business ourselves. It became clear to me that human oversight is the key for AI-enabled practices.
Finally, it is crucial to acknowledge the dependency of our business idea on Big Tech companies. It can be said that artificial intelligence is controlled by a few large tech firms, including Amazon, Google and Microsoft (Van der Vlist, Helmond & Ferrari, 2024). Almost all tech startups and established businesses that want to implement AI in their business operations will have to rely on the cloud services of Big Tech. This also became clear on the Demo Day, where I noticed that the vast majority of the working prototypes were built with the help of a few popular AI models. Our prototype was no exception, as it was developed using Anthropic’s AI model Claude. It is important to recognise that AI is heavily reliant on the cloud infrastructure provided by major technology companies, and it is crucial to consider the possible risks that come along with that reliance.
References
Pararius (2026, January 14). Affordable rental homes are disappearing from the market in the Netherlands. Retrieved from https://www.pararius.com/news/affordable-rental-homes-are-disappearing-from-the-market-in-the-netherlands
Renthaven (2026, August 17). Rentbird vs RentSlam vs Stekkies (2026): the full matrix, checked on 17 August. Retrieved from https://renthaven.nl/en/blog/rentbird-vs-rentslam-vs-stekkies
Sitanala, S. (2026, May 17). Huizenmarkt koelt af, maar we blijven overbieden: ‘Zet niet al je spaargeld in’. NU.nl. Retrieved from https://www.nu.nl/economie/6395948/huizenmarkt-koelt-af-maar-we-blijven-overbieden-zet-niet-al-je-spaargeld-in.html
Van der Vlist, F., Helmond, A., & Ferrari, F. (2024). Big AI: Cloud infrastructure dependence and the industrialisation of artificial intelligence. Big Data & Society, 11(1). https://doi.org/10.1177/20539517241232630
