Vigilance Cells
A screening layer ahead of a preliminary enquiry: the whole exported book read at once, ranked by corroboration, with the figures behind every flag attached.
A Chief Vigilance Officer works on complaints, source information and whatever the department can pull together in the time available. NIRNAE reads the whole exported book rather than a sample, and returns a ranked list of entities and transactions that merit verification — each with the rows behind it and the arithmetic that produced the flag, so a preliminary enquiry starts from figures rather than from an allegation alone.
Before a preliminary enquiry, to decide which of a year's transactions deserve scrutiny. During one, to test whether a specific pattern appears elsewhere in the same book. After one, to check whether the entities named recur across other schemes, districts or financial years. It is a screening layer ahead of the department's own process, never a substitute for it.
Payments split to remain below a sanction or delegation ceiling, vendors sharing an address, account, telephone or directors while bidding against one another, duplicate and non-measurable billing, year-end dumping and same-day sanction-to-payment, firms incorporated shortly before a large award, expenditure booked to a head that does not match its stated purpose, and value routed through intermediaries that performed no discernible work.
It does not verify that work was executed on the ground, establish intent, name any person as culpable or issue a finding of misconduct. It cannot see records the department has not supplied, holds no statutory authority, and has no access to any live treasury, banking or departmental system. Every output is an indicator for verification by the officers who do hold that authority.
On the department's written instruction, with the scope of records, the data-handling terms and the retention window agreed in advance. Typically one scheme, one district or one financial year of exported records, run alongside the existing process so the cell can judge the hit rate against what its officers already know.
Sample output
The figures below are from an illustrative sample ledger, shown so a cell can see the shape of the output before submitting anything. They are not live data and not drawn from any department.
Illustrative sample data. NIRNAE has no live access to any treasury, banking or departmental system, and holds no departmental record until one is formally submitted.
Transactions read
48,210
Whole exported book, not a sample
Entities flagged
126
Ranked by corroboration across agents
High-severity cases
19
Raised by three or more agents
Value under review
₹214 Cr
Sum of flagged transactions
Severity rises with the number of agents that independently corroborate a finding. A reviewer starts at the top of this distribution, not at row one of the ledger.
Each submitted export is mapped, merged and reconciled before the detection pass runs. Row counts and mapping status are reported back so the cell knows exactly what was examined.
Disclosure
Operating boundary. NIRNAE holds no live access to treasury, PFMS, banking or departmental systems. The department exports the records it wants examined and submits them. The product is in pre-deployment: engagements today are supervised pilots on exported or anonymised datasets, and every finding is an indicator for verification, not a determination. Full answers in the FAQ.
Leave a departmental email and we will arrange a briefing on a sample dataset.
Request a briefing