A menu for Chhattisgarh's AI Mission

Kisan Saathi

A WhatsApp line for the state's 25 lakh paddy farmers: where do I sell, when is my slot, has my money arrived, and who do I complain to.

What a farmer experiences. He messages the state number, or sends a voice note in Chhattisgarhi. He gets his token and slot, the queue at his centre, and the day his payment lands with the UPI reference. If something is wrong he logs it in the same chat and gets updates until it is fixed. Behind it, Markfed sees which centres will jam next week and which payments are running late, before farmers feel it.

Counts toward: 50 AI-enabled services, the agriculture pillar of Anjor Vision. Builds on: the existing token app and Markfed command centre. Needs from the state: Markfed data access and a named product owner.

No Family Left Out

The state works out what each family is entitled to and has not claimed, and messages them to claim it in the same chat.

What a family experiences. A WhatsApp message: "Your household qualifies for these three schemes you have not applied for. Reply to apply." They upload a photo of the document asked for and get a reference number. No trip, no agent, no form. The same rules engine quietly flags duplicate or ineligible claims before money moves, which protects the schemes from the fraud stories that have already surfaced.

Counts toward: AI in governance, Mahtari Vandan, Anjor Vision outcomes. Needs from the state: permission to join scheme databases, with the platform's privacy controls.

Jan Samadhan AI

Sushasan Tihar's intake and follow-through, run by AI every day of the year: every application read, routed, tracked, and closed with a call back to the citizen.

What a citizen experiences. At the camp, a worker photographs the paper form or records the request. The citizen gets a WhatsApp reference number on the spot, then updates as the case moves, then a message or call asking "was it solved?" A "no" reopens it. Officers get a queue with the right department already chosen and a drafted reply. The CM and collectors see a live league table of districts: received, resolved, overdue.

Counts toward: the Sushasan brand, 50 AI-enabled services, KPI-based reviews in Anjor Vision. Needs from the state: access to the campaign software through CHiPS, and a start early enough for the 2027 campaign.

Padhai Saathi

A personal tutor for every school subject, in Hindi and Chhattisgarhi, on WhatsApp and a light app. And English taught from Chhattisgarhi, which nobody offers today.

What a student experiences. She asks a maths question by text or voice note and gets it explained in her own language, at her level, with a practice question after. Ten minutes a day of English from Chhattisgarhi, the way language apps do it. It works on a cheap phone and offline in a lab. Teachers see who is stuck. The content comes from the state's own textbooks, generated with AI and checked by state teachers, so the cost of content is small. A child-safety guard sits between the student and the AI at all times.

Counts toward: 6 lakh students, 100 AI data labs, AI skilling pillar. Needs from the state: SCERT sign-off on content, teacher time for review, an education mandate to use it in schools.

Seva Setu on WhatsApp

All 520 Seva Setu services in one WhatsApp chat: find the right service, check you qualify, apply, pay the fee by UPI, and receive the certificate as a QR PDF. By voice note if you prefer.

What a citizen experiences. "I need an income certificate." The chat asks three questions, lists the documents, takes photos of them, takes the Rs 30 fee by UPI, and delivers the certificate to the same chat when it is signed. No kiosk, no agent, no reading a form. Seva Setu already has WhatsApp tracking and Bhashini text; this adds the conversation.

Counts toward: 50 AI-enabled services, Bhashini, AI in governance. Needs from the state: Seva Setu APIs through CHiPS, one rules expert per department.

Kaushal Saathi

Skills the way employers ask for them: companies post the roles they need, the platform generates the course in the local language, the learner certifies, the placement is tracked.

What a young person experiences. He sees real openings near him (the data park, Bhilai, mining contractors, hospitals), each with the exact course to qualify and the employer's promise to interview certified candidates. He learns on WhatsApp and the app in Chhattisgarhi, ten minutes at a time, and retrains whenever he wants without an application form. Employers get a pipeline they helped design.

Counts toward: 6 lakh students, employment pillar of Anjor Vision, AI skilling. Shares the engine with Padhai Saathi. Needs from the state: an employer council with named commitments and apprenticeship funding.

Satya Setu

Catches public-safety and fraud rumours as they spread and gets the state's answer out in the same language within two hours. Response, not censorship.

What a citizen experiences. A fake "your Mahtari Vandan payment is blocked, click here" message is doing the rounds. Within two hours the official clarification is on the state's WhatsApp channel, in Chhattisgarhi, and anyone who forwards the rumour to the state's tip line gets the answer back. The system watches public platforms (not private messages), works only on six categories (health, disaster, communal tension, scam and fraud, impersonation of government, safety in Maoist-affected areas), never orders removals, and publishes every clarification on a public log. Built on a national information-integrity platform we designed for a federal government.

Counts toward: awareness and outreach pillar, AI in governance. Needs from the state: Jansampark as owner, a written charter approved by the law department, two outside members on the review panel.

Chhattisgarh Language AI

Speech and translation AI for Chhattisgarhi, Gondi, and Halbi, built from recordings the state pays villagers in Bastar and Surguja to make. Released open, so every product above works in tribal languages.

What a villager experiences. A self-help group member records everyday speech on a phone app and is paid by UPI for each verified hour. Two thousand contributors, paid directly by the state. A year later, the paddy line, the grievance system, and the school tutor all understand Gondi. Today nothing does: national platforms cover Chhattisgarhi text and Gondi text only, and no speech at all.

Counts toward: centres of excellence, Bastar Anjor, contribution to Bhashini. Needs from the state: a university partner, Tribal Welfare facilitation, contributor payments in the state's name.

Sushasan Observatory

The data from Jan Samadhan and the other products turned into a live map of what every village asks for, how fast every district responds, and where a scheme is failing.

What the CM experiences. One screen: applications by gram panchayat, resolution time by district, the ten departments with the longest queues, and what will breach its deadline next week. A public version, the Sushasan Index, published monthly, lets citizens see how their district performs. It is also the KPI layer Anjor Vision 2047 asks for. Aggregate only, never individuals.

Counts toward: Anjor Vision KPI reviews, AI in governance. Needs from the state: the CM's decision on whether the public index is published.

Alert Chhattisgarh

Warnings that reach the right villages in the right language before the flood, the heatwave, the lightning storm, or the malaria season.

What a villager experiences. A WhatsApp message and a voice note in Gondi the evening before the river is expected to rise, with what to do and where to go. Targeted by gram panchayat, not broadcast to the whole state. Cheap, because it rides on the same number and channels as everything else.

Counts toward: AI in governance, Bastar Anjor. Needs from the state: access to the disaster and health alert feeds.

CG AI Platform

The shared foundation every item on this menu runs on: one state WhatsApp number, one language gateway, one set of AI models hosted in the state, one privacy and safety layer, one developer portal for startups.

What the state gets. A platform it owns, hosted in Nava Raipur, that any department can put a new service on in weeks. Every AI system on it is tested before launch and listed on a public register, which is how the "safe and responsible AI" pillar becomes real. The developer portal with sandbox data is how the state serves its 300 startups without a separate programme. Handed over to CHiPS with a trained team.

Counts toward: safe and responsible AI pillar, innovation ecosystem, the Nava Raipur centre of excellence. Needs from the state: a hosting decision and Bhashini access through MeitY. Required by every other item.

AI-era Security Layer

Security for the new attack surface AI creates: the WhatsApp and UPI front doors, the AI models, and the automated actions behind them, plugged into the security operations centre the state is building at Nava Raipur.

What the state gets. Detection tuned to attacks a normal SOC does not see: attempts to trick the assistant into leaking data or approving the wrong claim, poisoned documents, model abuse. Controls on what any AI system is allowed to do. Red-teaming of every state AI system before launch and every year after. This is our core business as an AI security company; the SOC itself is the state's.

Counts toward: safe and responsible AI pillar, the Nava Raipur SOC. Needs from the state: a decision on whether the IT or Home Department owns it.

Sarathi

An AI assistant on every government desk for drafting, rule lookup, minutes, and translation, with a three-level certificate earned by using it.

What an officer experiences. "Draft a reply to this RTI." "What does the 2024 order say about this scheme?" Answers cite the rule. Nothing leaves the state's platform. Short lessons inside the tool lead to AI User, AI Practitioner, and AI Nodal Officer certificates, which is how 1.5 lakh employees get trained without classrooms.

Counts toward: 1.5 lakh employees trained, AI nodal officers. Shares the engine with Padhai Saathi and Kaushal Saathi. Needs from the state: a training mandate.