Writing archive

AI will not remove managerial responsibility. It will move the bottleneck from producing work to judging, reviewing, and redesigning work.

Corporate employees do need GenAI skills, but prompt engineering is too narrow a starting point. The real skill is knowing where AI belongs in work.

Three Indian B-schools claim to be 'first AI-native.' The honest translation of the term, from how a16z and YC use it, exposes a contract problem most B-schools are avoiding.

Six months of building writing skills, audit pipelines, and self-correcting agents taught me one thing: the prompt is the wrong primitive. Here is the stack I switched to.

Most enterprises treat context engineering as a prompt or RAG problem. The real bottleneck is upstream: which source wins, which exception matters, who's accountable for the recommendation.

Research from Stanford, Harvard, and Carnegie Mellon points to a pattern leaders need to understand: AI can help less-experienced workers more when experts stay attached to old processes.

India's AI readiness debate cannot stop at infrastructure and talent. The harder question is whether users and organisations will pay for the tools and habits that make AI useful.

The second part of the IIM GenAI course argument: the market often pays for credential insurance, while real AI readiness depends on changed institutional practice.

IIM Bangalore's renamed analytics programme shows a wider problem in Indian executive education: the credential may be valuable, but AI readiness needs a different test.