Information Strategy – Project learning blog

For the AI Strategy Lab, my team(Team 22) developed Enerwise, an AI-powered platform that can continuously monitor household energy contracts, identify the eight most attractive alternatives, and explain why specific recommendations are suitable for each customer. At the beginning of the project we considered several ideas addressing different problems. Eventually, we decided to pursue my proposal, Enerwise. I believe the reason the idea resonated with the group was that we had all experienced difficulties with contracts in some form in the past. Specifically as international students in the Netherlands, choosing an energy contract can be quite confusing because of unfamiliar terminology, country-specific regulations and uncertainty about which contract conditions will actually matter. What initially seemed like a relatively straightforward comparison problem in the end taught me a lot about the relationship between AI, customer value, and business strategy.

My main contributions were developing the initial idea, working on the problem & opportunity analysis and the economics section. Since I proposed Enerwise, I already had some background knowledge of the problem, while my bachelors in Economics made the financial analysis a natural responsibility. This reflected the broader approach to dividing tasks: we assigned sections based on each member’s strengths and experience. I think this was one of our team’s biggest advantages. Everyone felt comfortable contributing within their area and our different technical and business backgrounds helped us make more informed decisions for all areas. Often someone’s expertise was valuable beyond their assigned slides because they could explain the reasoning behind decisions that affected the entire project. However the division of responsibilities also created challenges. While it allowed everyone to work independently and efficiently it sometimes made it difficult to understand the decisions made in other parts of the project. For example, I was not closely involved in the development of the prototype which meant for example for the AI cost I didn’t know what type of model the prototype would use for the explanations. While this was only a small issue it did make me realize that dividing tasks does not eliminate the need to collaborate, and makes consistent communication even more important.

As the project progressed I gradually took on a more coordinating role. I helped structure meetings, identify what we needed to discuss, set intermediate deadlines, and check with everyone on their progress. I also contacted the lecturer and TA when we needed clarification or feedback. We didn’t assign this role but for me it was important especially as we split up tasks having a clear overview of our progress was important to be able to keep on track. Alongside this I found myself challenging original assumptions and early results, which during this project I found to be an important task.

One of the most important examples of this was our initial idea focused on automatic switching for customers to better energy contracts. From a technological perspective, full automation seemed like the most attractive solution because it would minimize the effort required from customers. However I started questioning whether customers would actually be comfortable allowing an AI agent to make financial decisions on their behalf, especially when they might not fully understand how the technology works. This led us to think more about customer control and trust. Rather than assuming full automation would create maximum value we developed a solution that allows customers to review recommendations and understand the reasoning behind them. For me this was an important turning point in my understanding of AI strategy. Before the project i often associated technological advancements with better products. However, Enerwise showed me the most technologically advanced solutions will not necessarily be the ones customers value the most. And while an autonomous agent might be technically feasible and could bring value to customers, if those customers do not trust it enough to use it, the positive effect of it can not come into play. This reinforced the central idea of the course that firms should not simply implement some form of AI as this does not mean it will bring value to customers.

Another important learning came from looking at Enerwise’s competitive advantage. Initially we only focused on what the Ai agent could do, such as providing personalized recommendations. However soon we realized that the underlying technology would not be difficult for competitors to replicate. What helped us here to make sure Enerwise has a sustainable advantage was the AI canvas. Specifically Enerwise could learn more about consumers preferences, consumption patterns and switching decisions, therefore improving future recommendations. This created the possibility of a feedback loop in which continued usage improves service.

The economics section, which I worked on, also provided another unexpected learning experience. While I knew theoretically that information goods typically have very low marginal costs and high initial development costs, I had never actually applied it to a digital product. When researching the costs of AI generated explanations and data storage i was genuinely surprised by how inexpensive some of these activities could become especially at scale. Seeing the actual numbers made the concept of near-zero marginal costs much more concrete. At the same time I also developed a better understanding of how pricing for digital products differ from pricing physical products. I was more familiar with cost-based pricing, where often prices are determined by simply adding a margin to production costs. For Enerwise, however, the low marginal cost of an AI recommendation should primarily reflect the value customers receive such as potential savings and reduced effort. This helped me understand many course concepts better by relating them to actual business decisions.

Looking back at our collaboration, I think our different backgrounds were a major strength but our coordination could have been better. Two team members did not live in Rotterdam, which made meeting in person more difficult. Although online meetings were possible I personally found that exchanging ideas and making decisions was easier in person. Because we worked relatively independently, we sometimes had to spend additional time explaining decisions or aligning different sections. In future projects, I would still divide responsibilities according to our strengths, but I would introduce more frequent check-ins and shared milestones from the beginning. I would also make sure everyone understood the main decisions across the entire projects, even when they were not directly involved in that section.

Overall, the AI Strategy Lab helped me translate theoretical concepts into practical decisions. I gained a much clearer understanding of the economics of information goods, the limitations of AI as a source of competitive advantage, and the importance of communication when developing an integrated strategy. Most importantly, the project changed how I evaluate technological innovation. Instead of first asking what AI could automate, I would instead focus on the main problems customers face and in what way AI could deliver value and more importantly if customers would use the proposed solution. The most important lesson here was that AI does not create value simply by being implemented. Its value depends on how well it is integrated into a business and whether it addresses real customer needs.