Data shows that 80% of AI projects fail. Why such high 'mortality' rate in AI? While CRISP-DM is considered as the closest development methodology for Data Science projects, the industry and practitioners still lack a clear framework for execution of DS projects. In this talk, Gopalan will be introducing a conceptual framework to systematically take a Business case from ideation to deployment with clear decision criteria in each stage. This helps an AI practitioner to 'pause or pursue' an AI idea.