Part 2 · The Fragile Link

This blog is based on exploratory field engagement and is presented for descriptive and design purposes. It summarises reported experiences and recurring patterns in responses, and does not make inferential or causal claims.

After tracing why many remained excluded in Part 1, this part focuses on those who did attempt to navigate the process, people who reached the counter, found help, or encountered new friction once they were inside the system.

“I didn’t think it would work… but the community centre helped me apply for Ujjwala. Two months later, I got a gas connection. No more smoke in the kitchen, a quiet but profound change.”

Shilpa — Housekeeper, Dindoshi

“After my father passed, we got ₹20,000 through a scheme. It paid for rituals and debts I didn’t know existed until an NGO told me.”

Resident, Dahisar

Problems reported by applicants

24%Reported not understanding the process
20%Reported rejection
17%Reported an informal payment or bribe

Problems reported during application

Confusion, rejection, and informal payments appeared repeatedly in applicant accounts, while a smaller share described the process as smooth.

Field observation: Awareness alone did not guarantee smooth access. Navigation often still depended on intermediaries, local officers, and informal negotiation.

When treatment breaks dignity

Many respondents described the application process not only as difficult, but as emotionally degrading. Several accounts suggested that dismissive treatment and procedural ambiguity were themselves barriers.

“Nothing moves without some money changing hands.”

Anil — Driver who described paying a middleman for PDS access

“We’ll fill your form, but you’ll only get ₹400 out of ₹3,400.”

Cook, Mumbai

17% of respondents explicitly reported an informal payment or bribe. The true figure may be higher, given the sensitivity of the question.

Field observation: Dignity violations appeared closely tied to poor outcomes, often weakening trust before formal delivery could even begin.

When applications went through — and when they did not

Among men, responses reflected a split between successful and unsuccessful applications. Fewer women reported applying, but those who did often described success when guidance or accompaniment was available.

Reported outcome by gender

Men appeared split between success and rejection; women reported fewer outright rejections but were also less represented in the applicant pool.

Reported outcome by religion

Outcomes varied across religious groups, though subgroup sizes were small and should be read descriptively.

Reported outcome by caste

Reported outcomes appeared broadly similar across caste groups, though hidden frictions and perceptions of unfairness remained present.

Field observation: When the process felt respectful and legible, people were more likely to describe successful outcomes, but smaller frictions persisted across groups.

Justice — and injustice

Even respondents who eventually received benefits often described the broader system as uneven. Many believed that deserving people were still left out while others benefited through contacts or discretion.

Perceptions of fairness

A large majority of respondents described welfare distribution as unfair or inconsistent.

“It’s not what you know, it’s who you know.”

Field observation: Fairness emerged as a fragile perception. Success in one case did not necessarily produce trust in the system as a whole.

Reimagining welfare with AI

When technology was described not as a replacement for officials but as a fairness-checking layer, many respondents reacted positively. Support appeared strongest when explanation, oversight, and human assistance remained in place.

Belief that AI could improve justice

A large majority of respondents expressed the view that technology could improve fairness if it made decisions more transparent and accountable.

“We don’t want to remove our officials. We just want someone watching to make sure they’re fair.”

Resident, Mumbai

These conversations helped shape how SARAL was imagined: not as a replacement for human officials, but as a way to make decision processes more visible and accountable.

Field observation: Technology was welcomed most when it was understood as strengthening fairness and transparency rather than replacing human roles.
Parth Mody
Engineer & Data Scientist — Building AI for Governance
← Previous · Part 1 Next → Part 3