AI for government: from RFP to a system that actually runs

Government AI has a quiet failure mode. A pilot gets built, a demo goes well, a press note goes out, and then nothing reaches the citizen. The gap is rarely the model. It is everything around it: procurement, data, trust, and the unglamorous work of fitting into how an office already runs.

Start from the problem, not the technology

The strongest briefs we see do not ask for AI at all. They describe a problem: a queue that is too long, a language barrier at a counter, a backlog of documents no team can clear by hand. Naming the outcome first keeps the project honest, and it makes success something you can measure later.

Data residency and accountability are not optional

For public data, where it sits and who can reach it is a starting condition, not a feature request. We build to the DPDP Act of 2023, with Indian data residency, the option to run inside government infrastructure, and a human in the loop on any decision that affects a citizen's entitlement. Every step is logged, so a decision can be explained and audited later. A system that cannot show its working has no place in governance.

• ▸Indian data residency, on government or sovereign cloud where required.
• ▸A human in the loop on decisions that affect a citizen's rights or benefits.
• ▸Audit logs in plain language, ready for scrutiny.
• ▸Service in the citizen's own language, starting with Telugu, Hindi, and English.

Design for the officer, not just the citizen

An AI tool that adds clicks to an already busy officer's day gets quietly abandoned, however clever it is. The systems that stick remove steps instead. They draft the reply, sort the queue, flag the exception, and leave the final call with the person who is accountable for it. Adoption is a design problem, and it is usually the one that decides whether a project lives.

In government, the hard part is almost never the model. It is trust, data, and fitting into the day.

Done well, public-sector AI is not about replacing officers. It is about clearing the routine work that buries them, so people are free for the judgement, the empathy, and the calls only a person should make.