We had the idea of using a mobile image recognition-based tool for the automotive industry to identify specific parts, where there is a high number of very similar parts being produced. Soon enough, we realized that solely identifying parts is not sufficient. The automotive industry is also highly dependent on the traceability of parts in case of warranty, faulty productions to calculate risk, etc. Therefore, in the automotive industry, they are already using an identification system that focuses on specified serial numbers and not only part types. In this case, we concluded, our tool would only be useful for small car repair shops with low levels of digitalization.
In conclusion, we identified the problem, stakeholders involved and understood the process around it. We had proposed an advanced solution but realised that this solution failed to solve the problem for the initial intended stakeholder and they already had an alternative in place that solved their issue.
The secondary stakeholder that did have this problem was a far more fragmented market and was tougher to target, with a lower willingness to pay. Additionally, we felt that sooner or later, repair shops would adopt the digital process from the OEMs (Original Equipment Manufacturers) who supplied them. Thus, we decided to kill the project after the discovery phase.
By having this framework of questions to answer, we built a clear picture that highlighted the flaws of entering this market. Even though it sounds attractive to build a machine vision system that identifies parts and digitises this information for traceability and decision making support, it is, ultimately, a tech-driven solution that is looking for a problem to solve.