I posted on LinkedIn on 2026‑10‑10 about the ten “Build What Moves India” finalists and reflected on what it means when an AI agent says “done”. I examined ten prototype projects, with three described in detail: Samarthan, JanSewak, and RailSetu. All are simulated integrations, and I argued that trust and human control must decide adoption.
Before it submits something, I would want to review the details, know who will receive them and choose whether to proceed.
And “done” should come with something I can verify.
Adding an approval button is only useful if the person understands what they are approving.
- Samarthan explores preparing a cyber‑crime complaint from a voice message.
- JanSewak proposes guiding people through government websites and forms.
- RailSetu looks at options involving seat changes on the same train when the full route is wait‑listed.