An NBFC analyst reviewing a borrower's account before an EMI default may see transaction activity change well before the bureau record changes. A missed payment or higher credit utilisation appears in bureau data only after the event, leaving short-tenure and high-frequency lending products exposed during the gap.
AA-connected bank accounts provide a more current view. Their transaction history shows what has happened in the past thirty to sixty days, including changes in inflows, outflows, and activity levels. The practical task is converting that raw signal into a query a risk team can review.
What Transaction Velocity Actually Measures
Transaction velocity combines count and value: the number of inward, outward, or total transactions in a period, together with their aggregate value. For early warning, the important comparison is usually not a universal threshold. It is the change from that borrower's established baseline.
Consider a borrower whose normal monthly activity is fifty transactions and Rs. 80,000 in aggregate outflow, but whose past three weeks show twelve transactions and Rs. 18,000 in outflow. That is a material change in behaviour. Travel or a shift in salary accounts could explain it, but the same pattern can also accompany a period of repayment stress.
The baseline is more informative than an absolute cutoff. A Tier 3 micro-entrepreneur may normally make six transactions a month, so a fall to two does not automatically indicate concern. For a salaried metro borrower averaging forty monthly UPI credits and ten recurring debits, the same compression would carry more significance.
Patterns That Have Predictive Structure
Certain velocity changes recur in borrower accounts before an EMI miss or NPA event. These patterns are not causally definitive. They are recurring enough in credit review data to justify monitoring and contextual review.
One pattern is a decline in inward transactions. Salary credits or business inflows may stop, arrive late, or become irregular before repayment trouble appears. AA transaction narration can help distinguish a salary credit, such as an NEFT with an employer prefix, from an ad hoc inflow or a recurring credit from the same originator on the same day each month.
Another pattern is a rise in small outward transactions. Multiple ATM withdrawals or UPI payments made in close succession, especially during the last week before an EMI due date, differ from outflows that remain within historical norms. Such clustering can indicate an attempt to move funds from the account that a debit is expected to access.
A third pattern concerns the composition of inflows. A borrower who once received one large employer credit each month but now receives several smaller credits from different narration sources may have experienced an employment change. That altered income structure can make the next EMI payment less reliable.
Building the Query
For AA-sourced monitoring, the lending team needs to define three elements for each borrower: the look-back period used to establish the baseline, the observation period for current activity, and the deviation thresholds that create a review flag.
The baseline commonly draws on three to six months of transaction history collected at onboarding. The observation window is generally the most recent fifteen to thirty days. Thresholds should reflect the product. A consumer lending team may flag a 40 percent fall in inward transaction count against baseline, while a business loan team may assess the current account debit-credit ratio differently.
With structured data, the query is direct. For every monitored account, calculate inward transaction count and total value for the observation window, then compare both with the same account's three-month baseline. Flag an account when either measure crosses its threshold, and show the supporting transaction-level detail so the analyst can review the context before taking action.
The harder decision is threshold calibration. A rule that flags 30 percent of the portfolio in a month can overwhelm a small review team. The objective is a flag volume that the team can examine in a working day, tuned against the product's historical default profile and the resulting deviation thresholds.
When Velocity Signals Help, and When They Do Not
Velocity monitoring is most useful when the borrower has active digital banking behaviour, such as regular UPI use, NACH debits, and monthly salary credits. These accounts tend to have a steadier baseline, making departures easier to interpret.
The signal is weaker for borrowers who mainly use cash, maintain accounts at several banks, or have naturally variable activity, as seasonal businesses do. A decline in the monitored account may then indicate account switching rather than financial stress. It remains a review prompt, but it requires additional context before action.
Life events can also create false positives. A planned vacation, a salary-account change, or a short banking disruption may produce the same velocity decline as an early warning event. The workflow should therefore route flags to review instead of triggering automated action. The query narrows the accounts needing attention; it does not decide the response.
How AA Consent Affects Monitoring
Periodic monitoring under the AA framework needs a different consent arrangement from a one-time underwriting fetch. A portfolio monitor generally requires a periodic fetch consent, allowing transaction data to be retrieved at defined intervals, such as weekly or monthly, during the consent validity period.
The borrower must give explicit consent, covering the data types, accounts, and fetch frequency. Some lenders present this monitoring consent with underwriting consent at loan origination, while others manage it separately. FIU obligations still apply: each periodic fetch must be logged, and fetching must stop when consent is revoked, with the resulting monitoring gap handled in the process.
This creates a real operational requirement. AA access is more structured than earlier approaches, but responsibility for that access is also clearer. A lender building velocity monitoring needs the consent lifecycle and audit trail to be correct at origination, rather than adding them later to an existing portfolio.