About
AnnSaarthi — a guide through procurement
Ann means grain. Saarthi means guide. The platform is named for what it does: it walks alongside a farmer from the moment they decide to sell until the payment reaches their bank account.
Procurement season concentrates an enormous amount of activity into a few weeks. Centres open at dawn, farmers arrive with a tractor-trolley of grain, and everyone waits. The waiting is rarely caused by the weighing or the quality check themselves — those take minutes. It is caused by the absence of a schedule that anyone trusts, and by the fact that no one inside the yard can see how the queue is moving.
AnnSaarthi treats that as an information problem rather than a paperwork problem. A farmer books a time window at a specific centre. The platform estimates what the wait will actually look like in that window, and says so honestly — as a range, with its confidence, and with the factors behind it. On the day, the farmer can watch their token advance. After weighing, they can see the accepted quantity and the payment moving.
Centre staff get the same picture from the operational side: who is arriving, how fast counters are clearing, and where the day is heading before the rush lands. Administrators see every centre in the district at once, so capacity can be moved to where farmers are actually waiting.
The design principle throughout: reduce uncertainty, not just paperwork.
The journey
Who AnnSaarthi is for
Farmers
A mobile service with large touch targets, minimal typing and plain language. It answers when to come, where to go, how long it will take and where the money is.
Procurement centre staff
A dense operations console for tablets and desktops: queue control, counter allocation, digital weighing, quality capture and capacity management.
Administrators and officials
Regional intelligence — a map of centre load, waiting-time analytics, procurement and payment oversight, model health and demand forecasts.
Design principles
- 01
Answer the five questions
Every screen should say what is happening, what to do, when, how long it takes and what comes next.
- 02
Honest AI
Predictions are ranges with confidence and stated inputs. A model failure degrades gracefully; it never blocks a task.
- 03
Plain language
“Your slot is confirmed”, not “Booking status: CONFIRMED”. “AI confidence: 91%”, not “0.91”.
- 04
Built for weak networks
Actions queue locally and sync when the connection returns, with the pending count always visible.