A cross-sell recommendation that is wrong for the borrower is not just a missed sale — it is a relationship cost. A home loan borrower who receives a personal loan offer has been told, implicitly, that the institution does not know what kind of customer they are. A salaried borrower who receives an MSME working capital offer has been told the institution does not know how they earn. The Cross-Sell Specialist AI recommends the next best product by matching the borrower's profile — income type, existing product, financial trajectory, and most recent signals — to the product that is most likely both to be approved and to genuinely serve the borrower's next financial need.
The product matching framework: four inputs to one recommendation
The Cross-Sell AI builds a product recommendation from four inputs processed simultaneously. The first is the borrower's current product and their behaviour on it — a borrower who makes prepayments on a home loan may be building equity towards a LAP; one who has never missed an MSME EMI may be ready for a working capital enhancement. The second is the borrower's income and employment type — a salaried borrower's cross-sell path is different from a self-employed proprietor's, whose path differs from an MSME's. The third is the borrower's financial trajectory — improving CIBIL score, growing income credits, reducing FOIR — which determines which products the borrower is now eligible for that they may not have been at origination. The fourth is the triggered signals — the specific combination of signals that crossed the readiness threshold determines which product fits the moment, not just the profile.
The output is not a single product name — it is a ranked product recommendation with a fit score for each, a specific loan amount range that the borrower's current profile supports, and the one or two reasons why this product fits this borrower at this moment, expressed in terms specific to their situation.
Product recommendation for Kaveri Constructions: ranked matches
MSME Working Capital Top-Up · ₹15–20 lakh · 36 months
Kaveri Constructions has a 14-month track record of on-time repayment on their existing ₹28L MSME term loan. GST revenue grew 34% YoY. A new GST registration in Tamil Nadu was filed last month — the expansion is underway. The ₹15–20L working capital top-up gives the business the liquidity to fund the Tamil Nadu operations before revenues from the new territory come in. The EMI calculator use on Nov 12 suggests the proprietor is already modelling something in this range. Expected FOIR post-top-up: 61% — within the 65% cap for MSME borrowers.
MSME Term Loan (standalone) · ₹20–30 lakh · 48 months
If the working capital top-up is not the right product for the Tamil Nadu expansion (e.g., if the borrower needs longer-tenor capital for machinery or infrastructure rather than working capital), a new standalone MSME term loan at ₹20–30L for 48 months is the next match. The borrower's demonstrated repayment capacity on the existing loan supports this ticket. The longer tenor keeps the incremental EMI manageable at ~₹55,000/month.
Loan Against Property (LAP) · ₹40–60 lakh · Against business premises
Kaveri Constructions owns the business premises (declared in loan documentation). If the Tamil Nadu expansion succeeds over the next 6–12 months, a LAP against the business premises at ₹40–60L would be the optimal capital structure for a larger phase 2 investment — secured, lower rate than an MSME term loan, longer tenor. The Cross-Sell AI flags this as a future conversation, not an immediate offer.
The product-borrower fit matrix: what the Cross-Sell AI checks for each product
| Product | Borrower profile fit | Key eligibility check | Disqualifying signals |
|---|---|---|---|
| MSME working capital top-up | MSME borrower with ≥12 months on-time, revenue growing, new market or product line visible | Post-top-up FOIR ≤65% · Existing loan at ≥60% repaid · No DPD in last 6 months | GST filing irregularity · Revenue declining · Outstanding >85% of original |
| Home loan top-up | HL borrower with ≥18 months on-time, property value appreciated, or income significantly grown | Post-top-up LTV ≤75% · Income supports additional EMI · No DPD in last 12 months | DPD in last 12 months · LTV would exceed 75% · Property in disputed jurisdiction |
| Loan Against Property (LAP) | Borrower with clear property ownership, income to service, needing larger ticket than MSME term allows | Property valuation current · Clear title · LTV ≤60% · FOIR post-EMI ≤60% | Property in co-name without co-borrower · CIBIL <650 · DPD in last 6 months |
| Personal loan (top-up) | Salaried borrower with strong repayment history, CIBIL improvement, personal financial need | Salaried income confirmed · Post-EMI FOIR ≤50% · CIBIL ≥700 · No missed EMI ever | FOIR already above 40% · MSME borrower (not appropriate product) · Any DPD in last 6 months |
| Education loan (new) | Salaried borrower nearing home loan maturity, income strong, dependent child approaching higher education age (17–20) | Life stage signal · Income supports education EMI · Loan nearly closed = capacity freeing up | No dependent child in profile · Loan still in early stage |
The opportunity score panel: what the Cross-Sell AI shows the relationship manager
The recommendation that is right for the borrower and right for the institution is the only recommendation worth making
A cross-sell recommendation that produces a disbursement the borrower cannot service is a disbursement that will become an NPA in 6 months. The institution earned the commission and lost the relationship — and the net credit cost exceeds the commission earned. The Cross-Sell AI's product matching explicitly checks post-disbursement FOIR, post-disbursement LTV, and current CIBIL before recommending any product — not because the institution is conservative, but because a loan that fits the borrower is the only loan that stays performing. A 94% fit score means the product is right for the borrower. It does not mean the borrower will accept — but if they do, both parties benefit.