Rethinking the customer interview: What data says about fraud, collections, wasted time 

For decades, customer interviews were considered a best practice for validating income, confirming employment, and uncovering risk. They became part of the operational culture of lending. Although the industry, fraud, consumer behavior and staffing models have changed, interviews remain — often without evidence that they still deliver value. 

Across more than 12 million loan originations and 2 billion data points, one trend stands out. Customer interviews are declining in effectiveness though they consume increasing amounts of underwriting and servicing time. 

Executives are asking sharper questions about these interviews:  

• Who actually answers the queries? 
• Do they improve collections? 
• Do they reduce fraud exposure? 
• How much time are we wasting?  

Is there a better way? The data suggests the answer is yes. 

Decoding the decline of customer interviews 

Here are some of the key findings across lenders: 

• Contact rates continue to fall, often below 20% for certain borrower segments. 
• Interviews rarely prevent fraud because misrepresented applicants often provide corroborated but unreliable answers. 
• Collections impact is minimal because repayment behavior is tied to income stability, not a one-time interview.  
• The operational burden is significant, with thousands of hours spent on outreach that adds little value.  
• Executives must evaluate whether the cost aligns with the impact. Increasingly, it does not. 

Why customer interviews fail to detect modern fraud 

Fraud evolves faster than verbal validation. Manipulated documents often look legitimate, and borrowers with fabricated documents can easily pass an interview. 

Informed’s fraud intelligence identifies: 

• Template generated paystubs used across multiple dealers; 
• Edited bank statements reconstructed with AI tools; 
• Income patterns outside statistical norms across two billion data points; and 
• Metadata inconsistencies no human can detect 

A borrower can pass an interview, but cannot fool a model trained on 80 million validated outcomes. 

Does the interview improve collections? 

Data and collections feedback align on the same conclusion: Interviews do not change repayment behavior.

Borrowers who plan to pay do so regardless. Borrowers who plan to default often remain unreachable. 

When controlling for income accuracy and employment stability, interview presence shows no measurable effect on delinquency. 

Therefore, better collections outcomes start with better verification at origination, not more interviews. 

How much time are we wasting? 

In some lending shops, interviews account for: 

• Hundreds of hours in outbound attempts; 
• Up to 10% of underwriting time; 
• Rework when interviews expire; and 
• Additional staffing during peak volume. 

Time is the cost lenders can no longer ignore. 

A better, cheaper, more accurate alternative 

1. Documentary intelligence

Two billion data points and industrywide fraud signatures verify income and detect misrepresentation more accurately than verbal confirmation. 

2. Predictive validation without documents 

Models trained on 80 million loan outcomes validate borrower stability without calling or interviewing. 

3. Text-based workflows and CollectIQ 

Borrowers respond to text; they ignore voicemail. Texts with automated prompts, borrower friendly responses and auditable communication trails tend to reach more customers for lower costs and with faster turnaround. 

Interviews were built for a world that no longer exists; text-based workflows are built for the world we work in today. 

The executive mandate 

Executives should reassess the role of interviews through four questions: 

1. Are we catching fraud, or simply feeling like we are? 
2. Are we improving collections or adding friction? 
3. How much time are we burning? 
4. Can AI-driven verification and text workflows deliver better outcomes at lower cost? 

For most lenders, the answer will be clear: Verification should be data driven; collections should be automated and text-native; and interviews should be redesigned, reduced or retired. 

AI-driven verification and digital customer contact do more than replace the interview, they deliver the outcome lenders have always wanted: Better data, lower cost, less fraud and more time for work that moves a portfolio forward. 

Jessica Gonzalez is the vice president of customer success at Informed.IQ and has more than 15 years’ experience in the financial services industry, including tenures at Santander Consumer USA and Visa.   

Content sponsored by Informed.IQ