From Entitlement to Access: Making Welfare Services Simpler with AI on WhatsApp
A government scheme can exist on paper, have funding behind it and reach millions of beneficiaries — and yet remain difficult for a citizen to access. The problem is often not the absence of a programme. It is the distance between knowing you are entitled to something and actually being able to access it.
We started working with Indus Action; organisation that works on enabling sustainable access to social protection — education, livelihoods, and other public welfare schemes — for socio-economically vulnerable citizens.
The result: Samadhaan Saathi, an AI-powered WhatsApp chatbot built for the Rajasthan Department of Social Justice & Empowerment (SJED) in partnership with Indus Action. It brings welfare information, eligibility guidance, application support and status information directly to citizens through WhatsApp — while deliberately keeping AI away from decisions where accuracy and authority matter most.
The challenge wasn't a lack of information. It was access.
This is where we saw an opportunity. Could technology create a simple first layer of assistance between citizens and the complex systems that power welfare delivery?
Rather than introducing another application, login or digital interface, Samadhaan Saathi was designed around WhatsApp. The chatbot allows citizens to access information about:
Post-Matric Scholarship
Social Security Pension
Palanhaar
They can ask questions about eligibility, documents, applications and status — in Hindi, Hinglish or English, using the script they are comfortable with.
Our approach to AI: use it where it adds value
Building an AI system for citizen services comes with an important responsibility. A generative AI model can be excellent at understanding language and explaining complex information. But it should not be trusted to independently interpret authoritative government data, calculate entitlements or make decisions.
We built a Status Interpretation Engine that deterministically translates raw government statuses into predefined categories.
For example, a status such as: “PENDING WITH INSTITUTE” is first classified by the system. Only then does the AI explain what that status means to the citizen and what they may need to do next.
This led to one of the core principles behind Samadhaan Saathi:
The system decides. AI explains.
Building for trust, not just convenience
When technology deals with welfare and vulnerable citizens, convenience cannot come at the cost of trust.
Samadhaan Saathi therefore incorporates explicit privacy and language safeguards.
The chatbot does not ask citizens for Aadhaar numbers, bank details, OTPs or passwords.
It uses controlled vocabulary to keep scheme terminology consistent, mirrors the citizen's language and script, and follows a gender-neutral, non-casual communication style.
These choices reflect something fundamental about how we think about technology at Impactyaan:
good UX in the social sector is not just about making something easy to use. It is about making people feel safe using it.
From a 10-week build to 3 lakh+ interactions
Since launch, Samadhaan Saathi has crossed:
30,000+ active users
3,00,000+ interactions
The chatbot is helping citizens answer routine questions around eligibility, documents, applications and status without necessarily visiting an office or depending on an intermediary.
Beyond Samadhaan Saathi
The architecture we developed is not limited to Rajasthan's welfare schemes.
The same model — a content-managed knowledge layer, deterministic decision or status logic and an AI conversational layer — can support many other public-service journeys.
Health entitlements. Agricultural subsidies. Disaster relief. Grievance redressal. Education schemes.
Anywhere a citizen needs to understand
“Am I eligible?”,
“What do I need to do?”,
“What is happening with my application?” or
“What does this status mean?”,
there is an opportunity to make access simpler. For us, that is the larger opportunity. Samadhaan Saathi is one step in that direction — using technology to reduce friction, increase transparency and bring citizens closer to the services and entitlements they are already meant to receive.
A note from Impactyaan
Building a chatbot is rarely just about building a chatbot. The real work is understanding what people need, the questions they ask, and the systems and processes behind every interaction. At Impactyaan, we work with mission-driven organisations to turn complex programmes, information, and processes into simple, accessible digital experiences — bringing together data, design, and technology to build for people first.
If you’re thinking about building a chatbot, making a service easier to access, or exploring what technology could do for your programme, we’d love to hear from you. Even if it’s simply a conversation to share ideas and compare notes, write to us at contact@impactyaan.com.
Comments