Citizen AI is RightWalk’s long-term vision for building equitable, conversational access to public services, starting with India’s apprenticeship ecosystem.
This document is not an architecture specification.
It is a product, policy, and philosophy narrative that explains why the system exists, what we learned from the field, and how we think about AI in public service delivery.
In current deployments, this vision is instantiated in specific domains (e.g., NAPS). Those implementations may carry working product names, while Citizen AI remains the stable, long-term public framing.
India’s apprenticeship ecosystem holds enormous promise, yet for many young people it remains inaccessible.
The barriers are structural:
For years, RightWalk’s field teams acted as a “human API”, helping students navigate the NAPS portal step by step.
The core question that shaped this work was simple:
Can every young person access the guidance of a skilled mentor and a digital facilitator — through a simple WhatsApp conversation?
Early attempts at deterministic, form-driven automation failed under real-world conditions.
What the field taught us was fundamental:
Human conversations do not follow predefined workflows.
Any system that assumes they do will fail the people who need it most.
This insight became the foundation of Citizen AI:
Repeated field interactions revealed a universal truth:
Students don’t wake up wanting to complete a profile.
They wake up wanting to find work.
This led to a decisive shift in philosophy:
This was not just a UX improvement — it was an equity decision.
Citizen AI is guided by a small set of non-negotiable principles:
The system must adapt to what users want, not force users to adapt to system constraints.
Ask for the least information required, only when it becomes necessary.
AI should reduce anxiety, not create it.
Confusion is a system failure, not a user failure.
Public-sector AI must be explainable, reversible where possible, and accountable.
Human support is not a fallback — it is a first-class design component.
Citizen AI is not envisioned as a chatbot.
It is a conversational access layer for public systems:
This framing allows the system to:
Field deployments fundamentally shaped the vision:
Most importantly:
Field teams are co-designers, not downstream users.
Citizen AI aspires to deliver meaningful guidance without invasive data collection.
The vision is:
Personalization should feel helpful, not extractive.
Over time, the Citizen AI vision extends beyond a single workflow or portal:
Each expansion is guided by the same principles: accessibility, consent, and fidelity to public systems.
Citizen AI is ultimately envisioned as reusable public infrastructure:
The goal is not ownership of users, but stewardship of access.
Citizen AI began with a simple frustration:
public systems that work on paper but not for people.
It continues as a commitment:
This is not just a system.
It is a philosophy for how AI should serve citizens.