The question before the money moves
Pre-shipment purchase-order finance is high-yield and hard to underwrite: the lender advances against an order that has not yet been fulfilled. Before disbursement, a credit team is really answering one question in several parts — is this a real order, from a real buyer, to a borrower who can deliver it, that is not already financed somewhere else? Each part has evidence behind it, and the evidence varies in how strongly it stands up. The job of a good evidence pack is to surface every item and label it honestly — registry-verified, lender-side, issuer-confirmed, document-signed, self-declared, or missing.
AssureLocker focuses on deals from ₹30 lakh — aggregated across one or more POs — where the diligence is worth doing and the pack actually moves the decision. What follows is the checklist, and where each item’s evidence comes from.
1. Entity & authority
Who is the borrower, and is the person signing authorised to bind them? This is the foundation: GSTIN active status, the company/LLP registry record (MCA21), PAN consistency, and the authorised signatory. Names, addresses and bank details should align across sources. Most of this is registry-verifiable — checked live against an authoritative source — which is the strongest tier of evidence a lender can stand on.
This tier also does double duty on disbursal speed: verified entity existence, registered address and authorised-signatory evidence support a lender in streamlining the manual contact-point verification (CPV) that often holds a sanctioned loan for weeks. The lender keeps its CPV policy — but with far less manual field-checking, the gap between “approved” and “disbursed” narrows.
2. The order, and whether it is authentic
The purchase order is the basis of the financing. Beyond reading its value, the lender wants the supporting invoice to be a genuine, GST-registered document — the e-invoice IRN proves the invoice is registered and unique, which is the single most useful guard against the same invoice being financed twice. Where an IRN is present, the order’s authenticity is registry-verified; where it is only an uploaded PDF, it is self-declared, and the lender should treat it accordingly.
3. Did the goods move?
For a receivable to be real, the underlying goods generally have to move. The e-way bill records the consignment — the parties, the value, the vehicle, the route and the date — and, cross-read with the seller’s GST returns, it corroborates that a dispatch consistent with the invoice actually happened. This is a corroborated signal rather than a registry-verified one: no single source certifies the deal, but two independent records agreeing on the same movement is far harder to fabricate than either alone.
In pre-shipment finance the movement often has not happened yet — that is the point of the advance. Here the lender is reading the borrower’s track record of moving goods against similar prior orders, not proof for this one. A deal with no movement history at all is not automatically bad — a first-time exporter with a genuine order has to start somewhere — but it is a flag the lender should see plainly, because no exchange or central facility is vetting it on their behalf. The value of surfacing it is that the credit team prices and conditions the advance knowing exactly which parts rest on history and which on the order in hand.
4. Is it already financed elsewhere?
Double financing is the most expensive miss on a bilateral book: the same invoice or receivable pledged to two lenders, each believing it holds first claim. It is expensive precisely because it is invisible from inside one lender — your own systems show the deal is clean; the conflicting charge sits on someone else’s book. The authoritative check is a CERSAIcharge / receivables search, run on the lender’s own membership and credentials inside their boundary — a lender-sidesignal, never a shared one. AssureLocker does not hold the lender’s CERSAI credentials or run the search on their behalf; it surfaces the check as a required step and captures the result the lender records.
A charge search alone is not the whole picture. Combined with consent-pulled bank-cashflow aggregates (Account Aggregator, again lender-side, computed to aggregates rather than raw statements) and a related-party screen across borrower, buyer and any intermediaries, the pattern becomes legible. This is where a conflict engine earns its place: the failures a single lender cannot see alone — the same receivable financed twice, a buyer and supplier under common control, or a circular-trade ring passing the same goods between related parties to manufacture financeable paper — show up only when the parties are screened together, not deal by deal. Surfacing the pattern is the pack’s job; deciding what it means for the advance remains the lender’s.
5. Can the borrower actually fulfil it?
A real order to a borrower who cannot deliver is still a loss. Workforce (EPFO), GST-registered facilities, turnover trend and the subcontractor chain together indicate capacity for thisorder — not just that the business exists. Where the order is far larger than the borrower’s prior history with that buyer, that is a review item, not a rejection — but it should be on the page.
How the evidence stacks up
The point of labelling each check by tier is that they are not equally strong, and pretending they are is how underwriting goes wrong. A registry-verified item was checked live against an authoritative source and is the firmest ground a lender can stand on. A corroborated item is not certified by anyone, but two or more independent records agree — strong, though not conclusive. A lender-sideitem is run inside the lender’s own boundary on their credentials, because it must be. And a self-declared item — an uploaded PDF with no registry backing — is exactly as good as the word of the party that supplied it, which the lender should weigh accordingly. Below, each check mapped to what it proves, where the evidence comes from, and how far it stands up.
| Check | What it proves | Source | Evidence tier |
|---|---|---|---|
| Entity & authority | Borrower exists; the signatory can bind them | GSTIN status, MCA21, PAN | Registry-verified |
| Order authenticity | The invoice is a genuine, unique GST document | GST e-invoice IRN | Registry-verified (PDF-only = self-declared) |
| Goods movement | A dispatch consistent with the invoice occurred | E-way bill + GST returns | Corroborated |
| Fulfilment capacity | The borrower can deliver this order | EPFO, GST-registered facilities, turnover trend | Corroborated |
| Prior charge / duplicate | The receivable is not already pledged elsewhere | CERSAI charge / receivables search | Lender-side |
| Cashflow & related-party | The deal is real, and not circular or self-dealing | Account Aggregator aggregates, related-party screen | Lender-side |
What the pack hands the credit team
AssureLocker assembles these into an evidence-linked Risk Signals Pack: every signal labelled by evidence tier, a conflict engine for the patterns lenders lose most on, missing evidence surfaced rather than hidden, and an audit-ready trail. This is what AssureSignal for Purchase Order Financing is built to do — turn the six checks above into one review-ready pack a credit team can act on in minutes rather than weeks.
AssureLocker is a Technology Service Provider. It assembles evidence and surfaces signals; it does not lend, does not price or set an advance rate, does not hold funds, and does not decide the deal. Where a check must run on the lender’s own credentials — the CERSAI search, the Account Aggregator pull — it stays inside the lender’s boundary, and the pack records the result rather than performing it. It is signals, not a decision: the advance rate, the price and the sanction are the lender’s, made in the lender’s own system of record. The pack’s value is simply that the credit team can say yes, or no, faster and on firmer ground — and, because it is built on the same reusable DigiKYC / DigiKYB identity layer, the cost of the next check on the same party falls toward zero.