Learn how a Co-Lending Partner Reconciliation AI Agent reconciles disbursements, collections and loan-level shares daily, resolves breaks and keeps reporting aligned with RBI co-lending rules.
A Co-Lending Partner Reconciliation AI Agent is an AI system that reconciles disbursements, collections and loan-level shares between co-lending partners every day. It enforces the agreed split, resolves breaks and keeps reporting aligned with RBI's co-lending directions.
Co-lending has become a major growth engine in Indian lending. NBFCs bring origination reach and customer relationships, while banks bring lower-cost funds.
Each co-lent loan, however, lives in two sets of books. Every disbursement, EMI, prepayment and charge must be split correctly and recorded identically by both partners.
The agent is a loan-level reconciliation and control layer that sits between partners. It compares what the agreement requires with what each partner's systems show and drives every difference to resolution.
It connects the originating partner's loan system, the funding partner's loan system and the escrow bank account, with links to finance and reporting systems on both sides.
Co-lending operations, finance, treasury, partnership managers, risk, compliance and audit teams at both partners.
Co-lending only works if both partners trust each other's numbers. Reconciliation gaps erode that trust and create regulatory and financial risk.
RBI's Co-Lending Arrangements Directions, 2025, which took effect on 1 January 2026, extended co-lending beyond priority sector loans and set common rules. These include a minimum share of each loan to be retained by every partner, escrow-based fund flows, uniform asset classification and limits on default loss guarantees.
A single partnership can involve thousands of new loans and lakhs of EMIs each month. Spreadsheet reconciliation cannot keep up.
A missed collection allocation can make one partner show the borrower as overdue while the other does not. That affects provisioning, bureau reporting and collections.
Partners settle shares of collections, fees and service charges regularly. Unreconciled data leads to disputes and delayed settlements. Our guide on partner payout reconciliation for fintech lenders explains similar challenges in adjacent partnerships.
Lenders with multiple co-lending partners need a repeatable reconciliation model, not a custom spreadsheet for each one.
The agent ingests data from both partners and the escrow account, matches it at loan level against the agreement, classifies breaks and drives resolution.
The agent converts each co-lending agreement into rules: share ratio, partner rates, blended rate method, fee sharing, collection waterfall, settlement timelines and reporting responsibilities.
It reads disbursement, repayment and loan status files from both partners' loan systems and transactions from the escrow account.
Each loan is matched across both partners using loan identifiers, borrower details and amounts. Unmatched loans, such as a loan disbursed by one partner but missing from the other, are flagged first.
The agent checks that each partner's principal, interest, fees and collections follow the agreed split. Small rounding differences are separated from real breaks.
Every credit and debit in the escrow account is matched to loan-level transactions. Our guide on co-lending escrow reconciliation explains why this account is at the centre of co-lending control.
DPD and asset classification are compared for every borrower across both partners. Mismatches are escalated immediately because rules require uniform borrower-level classification.
Breaks are classified by type, cause and owner, such as timing difference, wrong split, missed posting or data mapping error, then routed to the right team at the right partner.
Each break is tracked to closure with ageing, and daily and monthly reconciliation reports go to both partners.
The agent makes co-lending partnerships more trustworthy, scalable and compliant.
The agent connects both partners' systems without forcing either to change its core platform.
Data from the originating NBFC's and the partner bank's loan systems is mapped into a common model.
Escrow account statements and transaction feeds are reconciled daily, alongside collections from channels such as UPI and NACH. Our guide on UPI collection reconciliation covers one of the fastest-growing collection channels.
Each disbursement is checked against both partners' records, as described in our guide on disbursement reconciliation.
Reconciled loan data feeds the Credit Bureau Reporting Data Quality AI Agent, so that bureau submissions for co-lent loans are consistent.
New partnerships start with clear data standards agreed during co-lending partner onboarding, while the co-lending eligibility filter decides which loans enter the arrangement.
Results show up in break volumes, settlement speed and compliance.
The agent supports all types of co-lending structures and products.
Home loans, loans against property, MSME loans, gold loans and personal loans co-lent between banks and NBFCs.
NBFCs working with several banks, each with different shares, rates and terms.
Ensuring prepayments and foreclosure proceeds are split correctly and closures happen in both books on the same date.
Applying the agreed order for allocating collections to fees, interest and principal for each partner.
Tracking any default loss guarantee given by the originating partner against RBI limits and invocation terms.
When co-lent loans are later sold or assigned, the Loan Pool Selection and Data Tape AI Agent uses reconciled ownership data to avoid selling loans that are already shared.
Reconciled data gives leaders a reliable view of co-lending performance.
Leaders can compare partnerships on volume, quality, break rates and settlement speed.
Accurate income data by partner informs future negotiations on share ratios, rates and fees.
Consistent DPD and asset classification support early warning and provisioning decisions.
Break trends show where operations or systems need investment before volumes grow.
Clean data builds confidence for banks to increase limits with trusted NBFC partners. Our article on AI agents in co-lending explores the wider opportunity.
Reconciliation depends on cooperation and clear agreements.
Both partners must agree on what data is shared, how often and in what format.
Ambiguous terms on fee sharing or collection waterfalls lead to recurring breaks. Agreements should be reviewed when automation starts.
Different day count conventions, rounding rules and posting times can create noise that must be separated from real breaks.
Some breaks can only be fixed by the partner. Governance must define timelines and escalation.
Borrower data shared between partners must follow consent, data protection law and the co-lending agreement.
Co-lending is moving toward real-time, shared data platforms.
Transactions will be matched as they happen rather than at the end of the day.
Industry formats for co-lending data exchange will reduce mapping work.
Partners may use shared loan-level records that both can see, reducing duplication.
As co-lending expands to more products under the new directions, reconciliation will need to handle more complex structures.
Joint governance makes reconciliation a shared responsibility rather than a source of friction.
Document data formats, cut-off times, break categories and resolution timelines agreed by both partners.
Assign break owners in operations, finance and IT at each partner.
Hold short daily break reviews and monthly governance meetings with senior representatives.
Define when unresolved breaks move to senior management at each partner.
Include reconciliation controls in internal audit plans at both partners.
It reconciles every co-lent loan between the partners on a daily basis, covering disbursements, each partner's principal share, interest at each partner's rate and the blended borrower rate, collections and their split, fees and charges, escrow account movements, prepayments, foreclosures, asset classification and outstanding balances in both partners' books.
RBI's Co-Lending Arrangements Directions, 2025, which took effect on 1 January 2026, replaced the earlier co-lending model framework for priority sector loans. The Directions apply to arrangements between banks and NBFCs, including HFCs, and cover matters such as minimum share retention by each partner, blended interest rates, escrow-based fund flows, uniform borrower-level asset classification, default loss guarantees, borrower disclosures and grievance redressal.
A break is any difference between what the co-lending agreement says should have happened and what the partners' systems show. Examples include a disbursement in one partner's books but not the other's, a collection split in the wrong ratio, interest computed at the wrong rate, an escrow credit not allocated, or different DPD or asset classification for the same borrower.
Collections, disbursements and prepayments happen every day. If breaks are left until month-end, they pile up, distort each partner's books, delay payouts and can lead to inconsistent asset classification and credit bureau reporting. Daily reconciliation keeps breaks small and easy to fix.
The agent compares DPD, SMA and NPA status for every borrower across both partners each day. Because co-lending rules require borrower-level asset classification to be uniform, any mismatch is flagged immediately, along with the cause, such as a collection posted on different dates or a missed instalment allocation.
Yes. Most co-lending partners run different systems. The agent maps data from both sides into a common loan-level model, reconciles them and produces exceptions in a format both partners can act on, without requiring either side to change its core system.
Yes. It produces loan-level and portfolio reports showing each partner's share, collections, outstanding balances and asset classification, along with a full audit trail of breaks and resolutions. These support internal audit, statutory audit, partner reviews and regulatory inspections.
Co-lending operations, finance and treasury, partnership managers, credit risk, compliance and internal audit teams at both the originating NBFC and the partner bank use the agent's outputs.
Reconcile disbursements, collections and shares with your partners daily. Talk to our AI experts about a Co-Lending Partner Reconciliation AI Agent for your co-lending business.
Ahmedabad
B-714, K P Epitome, near Dav International School, Makarba, Ahmedabad, Gujarat 380051
+91 99747 29554
Mumbai
C-20, G Block, WeWork, Enam Sambhav, Bandra-Kurla Complex, Mumbai, Maharashtra 400051
+91 99747 29554
Stockholm
Bäverbäcksgränd 10 12462 Bandhagen, Stockholm, Sweden.
+46 72789 9039

Lewes
16192 Coastal Highway, Lewes, Delaware 19958, USA