AiSewak
Rajasthan · Unified citizen helpline and grievance intake

बाईसा

Baisa

One number for every department in Rajasthan: the agent works out which department the problem belongs to, answers, routes, registers a grievance and closes with a satisfaction rating.

Languages
Hindi · English · Marwari and Mewari understood
Channels
Inbound and outbound phone · WhatsApp · Browser voice call
Try the live demo →

The problem

Rajasthan's citizen interface is not short of channels. Sampark 181, e-Mitra, PHED, RajSSP and the discoms all publish numbers. The citizen's first task is therefore to guess which door to knock on, before they have described the problem to anyone.

Guessing wrong is expensive. A household whose Jan Aadhaar is not updated, or whose certificate is missing, abandons the application somewhere between the first call and the counter, and no single system records that it happened.

Grievances and unmet demand then sit in departmental silos, so district-level insight reaches the Chief Minister's office on a quarterly cycle — after the block that was failing has failed for a quarter.

What AiSewak built

बाईसा inverts the routing burden. The citizen describes the problem in their own words; the agent detects the intent — health, pension, water, electricity, land, a certificate or a complaint — and does not require the citizen to name a department at any point.

The call flow is explicit and ordered. She opens as an AI assistant and locks the language. She works out the intent. She answers from the official knowledge base, and when another department is the correct one, she explains why and hands over its helpline, website, timings and document list rather than a transfer tone.

She then covers the two things a routing layer usually cannot: scheme guidance — benefits, eligibility, documents and where to apply — and grievance registration on Rajasthan Sampark 181 with a case reference the citizen can quote back. Distress is escalated to a human line rather than handled.

Every call closes with feedback: whether the issue was resolved, a one-to-five satisfaction rating and a free-text suggestion. That closing step is what turns a helpline into a measurement instrument, and it is the part most helplines drop first.

Coverage in the public demo spans Ayushman Aarogya (MAAY), Lado Protsahan, RajSSP pensions, Palanhar, NFSA ration entitlements, e-Mitra and Jan Aadhaar services, Sampark 181 complaints, and the emergency numbers 112, 108 and 1090.

How it runs

Conversational layer

An ElevenLabs Conversational AI agent, Hindi-first with a language lock and Marwari and Mewari comprehension, so a caller using regional vocabulary is understood without being asked to repeat themselves in standard Hindi.

Live official-source lookup

A server-side search tool scoped to Rajasthan state domains — rajasthan.gov.in, sso, jankalyan, e-Mitra, Apna Khata and Jan Aadhaar — alongside the central scheme portals, so a changed circular does not have to wait for a knowledge-base rebuild.

Second channel

A parallel WhatsApp helpline covers scheme eligibility, grievance filing with a photo, e-Mitra services and DBT status, for citizens who would rather type and attach than talk.

Guardrails

AI-disclosed on every call; never an officer, a minister or the Chief Minister. No full Aadhaar, bank account, OTP or password is ever requested. Live lookups are confined to Rajasthan and Government of India portals, and approval, amount and timeline are deferred to the department. Conversations are scored automatically for compliance.

What a department can replicate

  1. 1

    The published pilot shape is two districts — the demo names Jaipur plus one rural district — carrying the core helplines and flagship schemes with e-Mitra, RajSSP and Apna Khata deep links, so baseline resolution is measured before the routing table is extended.

  2. 2

    The rollout phase standardises both the department and scheme knowledge-base templates, adds a Sampark call-centre copilot for the humans already answering, and opens district dashboards for Collectors and Zilla Parishad CEOs.

  3. 3

    For any state running a single-number helpline, the reusable design is the ordering: detect intent before asking for a department, hand over contact details instead of a blind transfer, issue a case reference on every grievance, and close with a rating. The rating is what makes the rest auditable.

Scope the same pilot for your department

Two districts, a bounded knowledge base, a written list of what the agent will refuse to do, and a measurement step in the call itself. That is the shape every pilot above started from.

AiSewak (AI Sewak) is a Voxdonna company, made in India.

© 2026 Donna AI Labs Private Limited · CIN U62013DL2026PTC464877. All rights reserved.