Defining a scorecard for an AI or machine learning projects
Success of any project and more so for an AI or Machine learning based project depends on in depth preparation in defining the problem statement, maturity of the organization stating the outcomes and more importantly a scorecard or measures for the outcome. An important factor for an AI or machine learning project is also mapping the maturity of the organization's processes and IT/ data management systems currently in place.
In most cases adoption of any new technology starts with short proof of concept or a trial to learn and evaluate the risks and challenges for a subsequent full scale adoption. Most of these trials use an extracted set of data and built on discrete systems which are not part or integrated with the enterprise infrastructure. This first stage of experimentation is fine as standalone. It helps the team to learn the intricacies. The challenge kicks in when the experience from these trial projects is used to design the full scale adoption with all its organizational dynamics rolled in.
The most important thing for an enterprise is to map their organizational readiness/maturity for adoption of machine learning or AI solutions. They could map themselves to being a beginner organization, an aspiring organization or an organization which is experienced in these technologies or someone who is ready to scale these solutions across the organization. This exercise is usually the moment of truth to look inward , assess and then decide on the roadmap, investments and therefore a timeline when the full scale roll out can be achieved. When looking for a vendor or a service provider or a consultant it is important for the organization also to assess their ability and experience in guiding the organization through these transformations.
When the organization assess itself accurately on their current maturity it clarifies across the levels of the organization on defining a clear and balanced scorecard. It then allows them to decide,monitor and course correct the timelines, experimentation, develop specific focused solutions, upgrade and integrate their IT systems, data cohesiveness and business processes such that the full scale organizational transformation can be successfully achieved.