Reactivation Agent AI

Invoked when: loan closure event fires in LOS, or monthly scan identifies borrowers in the 6–18 month post-closure window without a new product

• –Identifies two categories of reactivation candidates: recent closures (loans that closed in the past 30–90 days where the borrower has not yet applied for a new product — early reactivation where the relationship is still warm) and aged lapses (borrowers 6–18 months post-closure who have not returned — later reactivation where the relationship needs rebuilding). The 6–18 month window is the primary target for reactivation because borrowers outside it have either chosen another lender as their primary (beyond 18 months) or are in the natural post-repayment pause that precedes new credit demand (under 6 months).
• –Monitors the bureau for re-enquiry signals on lapsed borrowers — a credit bureau enquiry from a lapsed borrower indicates that they are actively shopping for a new loan. This is the highest-priority reactivation signal, because it identifies a borrower who is in the market right now. A bureau re-enquiry on a lapsed borrower triggers an immediate reactivation outreach rather than waiting for the monthly scan cycle — timing is everything when a borrower is comparing options.
• –Scores lapsed borrowers by reactivation likelihood using their prior loan history: NPS score at closure (satisfied borrowers are more likely to return), repayment record (borrowers who paid consistently are better credit risks and receive better win-back offers), closure type (pre-closure borrowers had sufficient cash to repay early — a positive signal; write-off closures are excluded from reactivation), and time since closure (the probability of return declines with time — borrowers 6 months post-closure have higher reactivation potential than those 18 months out).

Invoked when: a lapsed borrower is confirmed as a reactivation candidate — offer is calibrated before the 21-day sequence begins

• –Calibrates the win-back offer by combining three inputs: the borrower's prior repayment record (a clean repayment history justifies a rate improvement on their next loan — this is the primary win-back lever), the current market rate environment from the Risk-Based Pricing Agent AI (the win-back offer should be at least as competitive as what the borrower would receive at origination today), and the borrower's estimated current creditworthiness (bureau data and the time elapsed since closure). The win-back offer is not a blanket discount — it is a rate that reflects the returning borrower's track record, calibrated to bring them back rather than simply to match a hypothetical competitor.
• –Checks win-back offer eligibility through the Credit Decision Agent AI before the offer is dispatched — a lapsed borrower whose bureau profile has deteriorated since their prior loan may not be eligible for the same product at the win-back rate. An offer dispatched without eligibility check creates the expectation of a loan the credit workflow may not approve; the eligibility check prevents this disappointment from occurring at the point of application.
• –For high-value lapsed borrowers (prior loan above ₹25 lakh), flags the reactivation candidate for relationship manager outreach in addition to the automated sequence. High-value borrowers respond better to personal contact than to automated sequences alone; the RM call provides the human relationship dimension that the sequence supports rather than replaces.

Invoked when: win-back offer is calibrated and eligibility confirmed — 21-day sequence begins on day 1

• –Runs a structured 7-step sequence over 21 days: Day 1 — relationship acknowledgment message (no offer, simply acknowledging the completed journey and expressing that LendingIQ values the relationship); Day 4 — soft product introduction (mentioning that a new offer is available, without stating specifics — curiosity before commitment); Day 7 — offer reveal (specific win-back offer with amount, rate, and a clear acceptance path); Day 10 — offer reminder with a different framing (addressing the most likely objection for the borrower's segment — rate, process simplicity, or speed); Day 14 — social proof message (how other returning borrowers have used their next loan); Day 18 — urgency message (offer validity window closing, specific expiry date); Day 21 — final message (offer expiry acknowledgment, open invitation to return when the time is right — no pressure, preserving the relationship for the next cycle).
• –Adapts the sequence channel per borrower — the channel on which the borrower most engaged during their prior loan tenure is the primary sequence channel for reactivation. A borrower who never opened emails but consistently read WhatsApp messages receives the reactivation sequence on WhatsApp. Channel mismatch between the sequence and the borrower's preferred channel is a common win-back failure mode; this agent routes to the historically responsive channel from the prior relationship data.
• –Tracks sequence engagement at each step — open rates, click rates, and application submissions. Where a borrower engages with an earlier step (clicks the Day 4 soft introduction) but does not convert, the subsequent steps are accelerated — a borrower who is clearly interested but hesitant receives the Day 7 offer reveal earlier than scheduled. Rigid 21-day sequences that do not respond to engagement signals waste the engagement window that the early steps have created.