Project Learning Blog – Emma

No ratings yet. At the beginning of this project, the brainstorming process was more challenging than I had expected. I initially approached it by thinking about interesting AI tools and possible applications, instead of identifying a problem that needed solving. As a result, many of our early ideas felt somewhat forced and lacked a clear purpose. The process became more productive once I shifted my attention towards everyday problems that AI could potentially address. At the time, my housemate was furnishing her room with IKEA products, and I had seen how difficult it was for her to find furniture that worked together in terms of dimensions, style and budget. I brought this example into our group discussion, which eventually became the starting point for Smart Space. Although the importance of starting with a problem rather than the technology had been discussed during the course, applying it in practice made its relevance even clearer.

One of our first challenges was determining whether our proposed solution was sufficiently different from existing alternatives. IKEA already offers several digital planning and visualization tools, which initially raised questions about the added value of our concept. However, examining these alternatives helped us identify a more specific opportunity. Rather than providing another tool for customers to design their own rooms, Smart Space would support the process of searching, comparing and combining products into a complete room proposal. For me, this demonstrated how AI can create value by changing an existing workflow rather than simply adding another technological feature.

Our team divided responsibilities while regularly discussing and reviewing each other’s work. My main individual responsibilities focused on the economics and monetization of Smart Space, as well as the governance and failure analysis.

Our original idea was to offer Smart Space for free, assuming that an easier furnishing process would encourage customers to purchase more IKEA products. Initially, this seemed logical, especially since IKEA’s core business is selling furniture rather than digital services. However, the course material on information economics and AI pricing highlighted several costs I had underestimated. Developing Smart Space would require substantial upfront investment, including integration with IKEA’s product, pricing and inventory systems. In addition, although AI services have relatively low marginal costs, these are not zero. Each generation or adjustment could create additional processing costs. This was particularly relevant because we wanted customers to continuously refine their room proposals. Unlimited usage could make the service more attractive, but also make operating costs less predictable, particularly with heavy users.

Another consideration emerged when I researched IKEA’s existing interior design services. IKEA already offers paid human advice, starting at approximately €30. Introducing a cheaper AI alternative could therefore cannibalize revenue from existing services. Until then, I had mainly considered the potential additional furniture sales. This made me realize that introducing AI also requires considering how it affects an organization’s existing business model.

Based on these considerations, I proposed charging €4.99 for unlimited AI advice for one room, potentially crediting this amount towards a qualifying IKEA purchase. I considered job-based pricing more appropriate than usage-based pricing because customers are primarily interested in completing a room design rather than the number of AI interactions required. Charging per prompt might discourage customers from experimenting with alternatives. The proposed structure could also create a distinction between free self-service planning, affordable AI assistance and more personalized human advice. This process helped me understand the relationship between value creation and value capture. Even if a service creates value for customers, this does not automatically mean the company can capture enough value to justify its investment.

Our interviews added another perspective to these decisions. Each team member interviewed four participants, resulting in 16 participants in total. Smart Space generally enabled participants to complete the room-planning task faster than IKEA’s existing website. However, the findings regarding willingness to pay were less straightforward. Some participants expected Smart Space to be free, while others were willing to pay approximately €5, particularly if the fee could be deducted from a furniture purchase. This was interesting because I had initially approached willingness to pay mainly as an estimate needed for our pricing model. The interviews demonstrated how differently customers can value the same service. Our proposed price therefore remained a hypothesis rather than a validated decision.

Although we divided responsibilities, regularly reviewing each other’s work helped us connect the different parts of the project. Small differences in how we conducted the interviews also showed me the importance of agreeing on consistency in our questions beforehand. Overall, I learned that good collaboration also requires continuously sharing our thoughts and decisions.

Overall, this project changed how I see the role of AI in business. Initially, I mainly saw AI as a way to improve a specific customer experience. However, working on Smart Space made me realize that successfully implementing AI often requires an organization to change their current ways of working. For IKEA, this would mean connecting the technology to existing systems, services and business processes, rather than simply adding a new tool. My main takeaway is that the value of a technological innovation depends not only on what the technology can do, but also on how well an organization can integrate it into its business.