Student’s role and contribution
Our project, ImpactIQ, focused on building an AI-supported solution for mid-sized construction contractors who face supplier disruptions. The goal was to help project managers evaluate disruptions by linking broken project information and weighing possible response options while leaving the final decision to the manager.
Throughout the project my contribution grew from brainstorming and writing to polishing the solution. One of the valuable parts of the experience was frequent communication with teammates. This helped us constantly challenge, broaden and enhance the idea instead of treating the project as a fixed thing from the start. My contribution began at the brainstorming stage. I arrived at the team meeting with several ideas for the project. After discussing the options we agreed that ImpactIQ was the strongest idea to develop further. During the proposal stage I focused on identifying the key assumptions behind ImpactIQ. This helped me understand clearly which business hypotheses needed to hold for the venture to succeed. I learned that assumptions are important because they make the uncertain parts of a business idea explicit. They push us to move from believing an idea will work to identifying what must be true, what could make the idea fail and what evidence would later be needed to validate it.
A key example in ImpactIQ was the assumption that disruption data from construction projects could be meaningfully compared. This assumption became especially important because the proposed data advantage depended on learning from disruptions across projects. Thus comparability, not just the amount of data, became a central assumption in ImpactIQ. This made me realise that data only becomes valuable when it is relevant to the decision being made. I had initially thought that having more historical data would automatically strengthen the recommendation, but I learned that poorly matched data can actually make the result less useful.
As the AI Strategy Lab progressed my main responsibility became governance, failure analysis and the future-market test. Through this work I learned that a strong AI solution is not about having more data or better technology but about whether it actually improves a meaningful decision. Governance also became more concrete to me: the system should not only be secure and human-supervised but should also recognise when there is not enough evidence to give reliable advice. Finally the future-market test made me understand that AI itself is unlikely to be a lasting advantage and that more defensible value comes from relevant data, workflow integration and trust.
Important decisions and iterations
One of the most important business decisions, in my view, was how we defined the characteristics of the dataset. Initially, I thought that the more data we had, the more effective the agent would be in giving the most appropriate recommendation to the contractor. However, I learned that the amount of data alone is not enough. For the recommendation to be effective, the historical cases need to be similar and comparable. Only when the data are relevant do the estimates become better. This was an important turning point in how I understood what makes the agent accurate and valuable. This finding also changed the design of the solution, as the agent should not provide a recommendation when there are not enough comparable historical cases. For me, this was an important example of evidence directly changing the way we designed the product.
Another important iteration was related to the AI capabilities themselves. As similar AI capabilities become widely available and easier to replicate, the focus shifts toward assets that are more difficult to copy, such as relevant historical data, workflow integration, and human judgment. These elements can make the solution more valuable and also increase trust in how the recommendations are generated. This shifted our thinking from treating the AI capability itself as the main advantage to focusing more on the data, integration and trust around it.
What surprised or failed
What surprised me the most was that the value of AI is not only in its ability to make recommendations. I was especially surprised that AI can also recognise when the available data is not strong enough to support a reliable recommendation. This made me realise that reliability also comes from knowing and communicating the limits of AI, not only from making the system more capable.
How feedback and evidence changed the work
In my opinion, the feedback from the lecturer and my teammates were the most crucial for developing the project. It helped us challenge some of our initial assumptions and focus more on the actual problem we were trying to solve, rather than becoming too focused on the AI solution itself. The main point of the feedback was focusing more on whether each part of the idea was really creating business and strategic value. Also, the evidence from our testing played an important role as I learnt that our initial assumptions needed to be refined rather than simply confirmed. For instance, the feedback helped us focus less on the general capabilities of AI and more on whether the recommendation actually improved the decision of the project manager.
The importance of collaboration
One of the most crucial parts of this project, in my opinion, was teamwork. Different perspectives can improve an idea because they allow it to be looked at from technical, business, pricing and user points of view. In our case, frequent discussions and meetings helped us brainstorm, challenge each other’s thinking and analyse different parts of the project in more depth. Instead of simply dividing the assignment into separate individual tasks, we were able to build on each other’s work and make the project more iterative.
Collaboration also helped us challenge assumptions that initially seemed obvious but still had important gaps. Having teammates question these ideas pushed me to explain my reasoning more clearly and make sure that my contribution remained aligned with the overall vision and goal of the project. Through this process, I strengthened my critical thinking and learned how valuable it is to both give and receive constructive feedback.
How the project changed my understanding of AI and business strategy
A strong AI strategy starts with the problem, not the technology. Through this project I learned that AI is not a tool that can automatically solve every business problem. It is important to understand the problem clearly and identify the key business assumptions before deciding how AI can create value.
I also learned that the strategic value of data depends on its quality, relevance and comparability rather than simply its volume. Testing the assumptions behind the business model early was therefore especially important because the evidence can change the direction of the solution. Finally the project made me understand that as basic AI capabilities become widely available competitive value will increasingly shift away from the technology itself and toward assets that are harder to replicate such as relevant data, trust, workflow integration and other scarce complements.
Overall this project was very valuable to me because it changed the way I approach AI ideas. I now focus first on the business problem, the assumptions that need to hold and the evidence needed to support them before focusing on the technology itself.
