Point-of-Sale Fraud in Consumer Durable Loans: Store Signals
Last reviewed: November 2026
Point-of-sale fraud in consumer durable loans is fraud committed at or through the retail store where a phone, appliance or electronic product is financed, often with the store staff's help. Because each loan looks normal on its own, most lenders find it only when the first instalments bounce, by which time the store has sold dozens more.
This guide is for chief risk officers (CROs) and heads of fraud control at non-banking financial companies (NBFCs), banks and fintech lenders that finance consumer durables at retail stores. It covers six store-level fraud patterns, a seven-step detection workflow, the store risk signals to watch and the metrics that show the controls are working.
Key Takeaways
- Point-of-sale fraud is a store-level pattern, not an application-level one, so it must be detected by store, salesperson and device.
- The store receives the disbursal, which makes it the main beneficiary when collusion happens.
- Detection lag drives the loss: every week a colluding store keeps selling adds bad loans.
- Good bureau scores do not protect you, because fraud often uses real, creditworthy identities.
What Is Point-of-Sale Fraud in Consumer Durable Loans?
Point-of-sale fraud in consumer durable loans covers any scheme where the store, its staff or people working with them obtain loans that will not be repaid as intended. The customer may be a victim, a willing participant or entirely fictional.
Vetting stores before they go live is covered in our guide to merchant onboarding for consumer durable loans. The vehicle-finance version of the same problem is covered in our guide to dealer fraud in vehicle loans; this guide covers detection after a store is live.
Which Store-Level Fraud Patterns Should Lenders Know?
Lenders should know six store-level fraud patterns: identity misuse, ghost sales, invoice inflation, faked down payments, device flipping and promoter gaming. Each leaves a different data trail.
| Pattern | How It Works | Data Trail |
|---|---|---|
| Identity misuse | Staff reuse a walk-in customer's Know Your Customer (KYC) documents for extra loans | Same device or photo across applicants |
| Ghost sale or cash-out | No product leaves the store; borrower gets cash minus a cut | No delivery confirmation, early default |
| Invoice inflation or model swap | Higher model invoiced than delivered | Serial or International Mobile Equipment Identity (IMEI) number mismatch |
| Faked down payment | Store pays or fakes the borrower's down payment | Down payment receipts with no borrower trail |
| Device flipping | Product resold at once, often back to the store | First payment default, same serials reappear |
| Promoter gaming | In-store promoters steer applications around approval rules | Approval spikes after rule changes |
Why Does Point-of-Sale Fraud Go Undetected?
Point-of-sale fraud goes undetected for five reasons: application-level review, reliance on bureau scores, no product verification, volume-based store incentives and slow feedback from collections. Each has a prevention check.
| Cause | Effect | Prevention Check |
|---|---|---|
| Each application reviewed alone | Store patterns invisible | Store-level scoring every day |
| Bureau score treated as fraud check | Real identities pass | Separate fraud signals from credit signals |
| Product never verified | Ghost sales and model swaps pass | Serial or IMEI capture and match |
| Stores rewarded on volume | Bad stores grow fastest | Incentives tied to early repayment |
| Collections data reaches risk late | Months of bad loans pile up | Weekly first payment default by store |
Why Is the Store the Main Beneficiary?
The store is the main beneficiary because, in most consumer durable loans, the money goes to the store, not the borrower. According to the RBI (Non-Banking Financial Companies – Credit Facilities) Directions, 2025 (RBI/DOR/2025-26/347, 28 November 2025), loans for a specific end use may be disbursed directly into the end-beneficiary's bank account.
So in a ghost sale, the store is paid in full for a product it never delivered. Merchant subvention adds another lever, which our guide to subvention scheme management explains.
What Is the Right Workflow for Point-of-Sale Fraud Detection?
The right workflow for point-of-sale fraud detection has seven steps: tier stores, capture evidence at sale, verify identity and account, verify the product, confirm delivery, monitor store performance and act fast. Steps 2 to 5 happen before or just after disbursal.
- Tier stores by age, volume and history, with tighter rules for new stores.
- Capture live photo, device data and location at the point of sale.
- Verify identity and the repayment bank account.
- Verify the product through serial or IMEI number and invoice.
- Confirm delivery with the borrower directly, not through the store.
- Monitor store, salesperson and device patterns daily.
- Act by limiting, suspending or exiting stores, and recover where possible.
Steps 1 and 2: What Should Be Captured at the Point of Sale?
Capture a live photo of the applicant, the device used to fill the application, the location of the session and the salesperson's ID. These four items turn each loan into data that can be linked across stores.
Repeat photos, devices and contact details are the clearest sign of staff-filled applications. Our guide to loan application deduplication explains how to match them reliably.
Steps 3 and 4: How Do You Verify the Borrower and the Product?
Verify the borrower's repayment account belongs to them, and verify the product through its serial or IMEI number against the invoice. A serial number that has already been financed, or that does not match the invoiced model, stops the loan.
Bank account checks are covered in our guide to penny drop bank account verification. Shared fraud databases add a cross-lender view, as our guide to Hunter fraud check integration explains.
Steps 5 to 7: How Fast Should Lenders Act on a Suspicious Store?
Lenders should act on a suspicious store within days, not at month-end review. Confirm delivery directly with a sample of the store's borrowers, and limit new approvals while the review runs.
Third parties can be held to account. According to the RBI (Non-Banking Financial Companies – Fraud Risk Management) Directions, 2026 (RBI/DoS/2026-27/463, 31 July 2026), show cause notices extend to third-party providers, and NBFCs need a process that gives them at least 21 days to reply.
Which Store Risk Signals Should Lenders Monitor?
Lenders should monitor eight store risk signals, each compared with similar stores in the same city and category.
| Store Risk Signal | What It Suggests |
|---|---|
| First payment default well above peer stores | Ghost sales, identity misuse or flipping |
| Many applications from one device or mobile number | Staff filling applications |
| Serial or IMEI numbers repeated or mismatched | Invoice fraud or model swaps |
| Applicants from far-away pincodes | Imported identities |
| Surge of maximum-ticket loans near month-end | Target-driven or staged sales |
| Approval rate jump after a rule change | Promoters gaming rules |
| Repayment accounts shared across borrowers | Mule accounts |
| Low delivery confirmation response | Products not delivered |
Synthetic and borrowed identities often sit behind these signals. Our guide to synthetic identity fraud detection explains how they are built.
How many weeks does it take your team to stop a store that is selling bad loans?
Talk to Our AI Development Specialists
Visit Digiqt to see how a custom AI agent can score stores daily and flag collusion early.
How Much Does Detection Lag Cost in Point-of-Sale Fraud?
Detection lag can triple losses from point-of-sale fraud in consumer durable loans, because colluding stores keep selling until they are stopped. In this illustrative scenario, a lender finances consumer durables at 2,000 stores, with an average loan of ₹25,000.
Assume 40 stores collude, each producing 60 bad loans a month, and 80% of each bad loan is lost. Compare a three-month detection lag with a one-month lag.
| Measure (Illustrative) | Three-Month Lag | One-Month Lag |
|---|---|---|
| Bad loans before stores are stopped | 7,200 | 2,400 |
| Amount disbursed on bad loans | ₹18 crore | ₹6 crore |
| Expected loss at 80% | ₹14.4 crore | ₹4.8 crore |
The lesson for leadership: detection accuracy matters, but detection speed matters more. Signals available on day one, such as shared devices and missing delivery confirmations, beat waiting for the first instalment to bounce.
Where Does Custom AI Agent Development Fit in Point-of-Sale Fraud?
Digiqt's custom AI agent development services include artificial intelligence (AI) and data strategy, minimum viable product (MVP) and proof of concept (POC) development, custom machine learning (ML) modelling, business process automation, and explainable AI and data insights. Engagements run through a dedicated development team.
For point-of-sale fraud, a custom AI agent could be built to:
- Score every store daily on early repayment, approval patterns and peer comparison.
- Link applications through shared devices, photos, phones and bank accounts across stores.
- Check serial and IMEI numbers against earlier loans and invoiced models.
- Explain each flag in plain terms, so fraud teams can act and stores can be told why.
Digiqt's process runs from a discovery and business understanding sprint, through problem statement analysis, feasibility study, model selection and dataset assessment, to model validation and deployment with continuous improvement. Start with a proof of concept on six months of your own store data, including stores you already closed for fraud.
Digiqt's fraud ring detection AI agent and loan stacking detection AI agent show how linked-entity patterns can be found in lending data.
Which Response Fits Each Store Risk Level?
The right response depends on how strong the evidence is and how long the store has been live.
| Store Situation | Recommended Approach |
|---|---|
| New store, first 90 days | Lower limits, full delivery confirmation |
| First payment default twice the peer rate | Sample delivery calls, cap approvals |
| Shared devices or accounts confirmed | Pause new loans, review every recent file |
| Collusion confirmed | Exit store, show cause process, recover subvention |
| Chain store with many counters | Score each counter and salesperson separately |
| Promoter-run counter | Rotate promoters, monitor their approvals |
Which Metrics Show Point-of-Sale Fraud Controls Are Working?
Five metrics show whether point-of-sale fraud controls are working: first payment default by store, non-starter rate, delivery confirmation rate, shared-attribute rate and time to suspend. Review them weekly by store, city and salesperson.
| Metric | Formula | What a Weak Number Signals |
|---|---|---|
| First payment default by store | Loans missing first instalment ÷ loans due, per store | Collusion or poor store quality |
| Non-starter rate | Loans with no instalment paid in 90 days ÷ loans disbursed | Loans arranged to default |
| Delivery confirmation rate | Borrowers confirming delivery ÷ borrowers contacted | Ghost sales |
| Shared-attribute rate | Applications sharing device, phone or account ÷ applications | Staff-filled applications |
| Time to suspend | Days from first flag to store suspension | Losses growing while teams review |
Which Mistakes Let Point-of-Sale Fraud Spread?
Four mistakes let point-of-sale fraud in consumer durable loans spread, and each is a policy choice.
- Judging stores on volume. The fastest-growing store may be the most dangerous.
- Treating good bureau scores as clean. Victims of identity misuse often have strong credit.
- Paying store incentives before the first instalment clears. That pays fraud in advance.
- Ignoring staff movement. Salespeople who move between stores take their methods with them.
What Is Changing Next?
New loans now show up in bureau data within days. According to the RBI (NBFC – Credit Information Reporting) Directions, 2025, as amended on 4 December 2025, lenders have reported new accounts and changes as on the 9th, 16th, 23rd and last day of each month since 1 July 2026.
That narrows the window in which one identity can be used for loans at several stores and lenders. Fraud teams should pull fresh bureau data before second and later loans, and add multi-lender stacking to their store risk signals.
Want to stop colluding stores in days instead of months?
Talk to Our AI Development Specialists
Visit Digiqt to see how custom AI agent development can help detect point-of-sale fraud across your store network.
Sources
- RBI: Non-Banking Financial Companies – Credit Facilities Directions, 2025 (RBI/DOR/2025-26/347, 28 November 2025)
- RBI: Non-Banking Financial Companies – Fraud Risk Management Directions, 2026 (RBI/DoS/2026-27/463, 31 July 2026)
- RBI: Non-Banking Financial Companies – Credit Information Reporting Directions, 2025 (RBI/DOR/2025-26/349)
- Digiqt: Custom AI Software Development



