India AI is a decision-intelligence problem

IThe wrong starting point

India does not need more AI demos that impress a room for five minutes and then vanish from the workflow.

The harder question is whether AI can fit inside the places where consequential work already happens: files, briefs, sector reviews, ministry dashboards, board notes, procurement decisions, crisis rooms, cabinet-facing summaries, and the everyday judgement loops of serious institutions.

That is a different problem from asking whether India should build models, buy compute, or launch another chatbot. Those questions matter. But they are not the full story.

The India AI question is really this: can we turn intelligence into a usable public and institutional surface?

IIThe rooms matter

Large public systems are not run from blank prompt boxes. They are run through inherited formats: notes, presentations, tables, policy summaries, meeting briefs, compliance trails, consultations, and the judgement of people who know the system.

That is why generic AI adoption advice often feels too thin for government and regulated sectors. It talks about productivity in the abstract. It rarely understands how authority, accountability, evidence, precedent, confidentiality, and institutional memory actually shape decisions.

In India, this matters even more because the scale is unforgiving. Energy security, education, infrastructure, social development, diplomacy, technology policy, and public finance all sit inside systems where a small misunderstanding can travel very far.

AI that ignores those rooms will stay decorative. AI that understands those rooms can become infrastructure.

IIIThe missing layer

The missing layer is decision intelligence: a surface that helps a person understand what changed, what matters, what evidence supports it, what options exist, and what the next decision requires.

That layer is not just a model. It is not just a dashboard. It is a product discipline that combines retrieval, explanation, structured data, workflows, institutional context, user interface design, and judgement about what the decision-maker actually needs first.

This is where India has a very large opportunity. We have dense institutions, enormous public datasets, world-class technical talent, complex sectors, and a growing appetite for AI. What we do not yet have enough of is the connective product layer that makes AI useful inside public decision-making.

IVEnergy is the testbed

Energy is a good place to see the problem clearly. The sector is strategic, technical, political, financial, and public-facing all at once. Crude prices, refining, gas infrastructure, marketing companies, taxes, logistics, geopolitics, and consumer prices are connected, but they are rarely visible as one system.

That is why I built Sanjaya: a public oil and gas intelligence surface that makes the Indian energy system easier to navigate. It is not the final answer. It is proof of direction.

The same logic sits behind Sanket, my browser-side security posture check. Serious institutions need readable signals, not just raw technical output. A system that cannot explain risk clearly does not travel far enough inside the organisation.

The subject changes. The underlying move stays the same: make the system legible, then build from there.

VWhat I am building

My lane is India AI for public decision-making: AI governance, energy security, institutional workflows, and command surfaces for domains where the cost of confusion is high.

That lane joins parts of my background that usually sit apart: physics training, HR and organisational understanding, public policy, ministerial writing, government workflows, energy-sector exposure, and the ability to build AI systems directly.

I am interested in the point where a serious institution asks: what changed, what matters, what are the choices, what should we watch, and how do we explain this cleanly to the person who must decide?

That is the work behind this site, the public explainers, Sanjaya, Sanket, and the private command-surface experiments that sit beneath them.

VIThe category

India AI should not be reduced to frontier-model nationalism or shallow automation. The larger opportunity is to build AI that respects Indian institutional reality and improves the quality of decisions at scale.

That means systems that can sit on top of public data, sector knowledge, government formats, human judgement, and accountability. It means AI that can explain itself well enough to be used by people who cannot afford vague magic.

This is the category I want to help define: AI for India-scale decisions.

Not just demos. Decision intelligence.