Voice fraud is now one of retail’s fastest-growing threats, and fraudsters are winning. A new survey by Modulate and Retail Dive’s Studio by Informa TechTarget found that eight in 10 retail and finance leaders faced moderately to highly sophisticated voice attacks in the past year. Tactics are evolving faster than most organizations can adapt.

Armed with AI-powered tools available on the dark web for as little as $20, scammers can now clone voices, impersonate customers and employees, and manipulate call center agents with alarming precision. More than half of surveyed organizations report that individual fraud incidents cost between $5,001 and $25,000 each, with some cases climbing far higher. 

As retailers rush to strengthen defenses, many are finding conventional security measures are failing on two fronts. Traditional tools are still letting fraudsters through, while simultaneously driving away legitimate customers. Survey data reveals where losses hit hardest, why traditional defenses are failing, and how retail leaders are adapting before competitors do.

The Customer Experience Tax

The deepest irony in retail fraud prevention is that the security measures designed to protect customers are inadvertently driving them away. According to the survey, 44% of leaders cite customer complaints about verification processes as their top fraud-related issue, ranking higher even than concerns about detection gaps or attack sophistication. The cure is creating its own damage.

The friction ripples through operations: longer call handle times, customers abandoning calls mid-process, increased callbacks, and agent burnout. Each friction point chips away at customer lifetime value in ways that rarely show up on a fraud loss report, but hit just as hard.

Retail is uniquely exposed here. Unlike banking, where customers may tolerate friction for the sake of security, shoppers have infinite alternatives and will simply leave. When verification drives away a shopper over a $100 transaction, retailers may be forfeiting $10,000 or more in lifetime value, referrals, and social proof.

“You see the best outcomes when you layer different forms of verification without requiring the customer to actively participate in each one,” says Mike Pappas, CEO of Modulate. Visible security theater creates friction; invisible, intelligent monitoring preserves the experience.

The Operational Drain Few Are Tracking

Beyond the dollars lost to fraud itself, other hidden costs erode margins from the inside. It shows up in investigation hours, staff retraining cycles, and call volume spikes that have nothing to do with customers actually buying things.

The survey data makes the labor burden concrete: 71% of respondents say their organizations spend at least 51 hours annually investigating suspected voice fraud incidents, and 18% report spending between 201 and 500 hours. For an enterprise fraud team of 20 investigators at a median cost of $50/hour, that 201-500-hour cohort represents $201,000 to $500,000 in annual investigation labor alone – before factoring in escalations, retraining, and opportunity cost. Meanwhile, 39% report higher call volumes driven specifically by fraud-related inquiries, such as customers calling back after failed verifications, legitimate users locked out of accounts, and fraud victims seeking resolution.

Training adds another layer of persistent costs, with 38% of respondents citing rising expenses to keep agents current on evolving attacks. For retail specifically, where delivery fraud, return abuse, loyalty program exploitation, and gift card scams all exploit voice channels, the surface area is enormous.

In retail, even a 1-2% increase in operational costs from fraud management can wipe out an entire year’s profit improvement. These losses rarely surface cleanly in the fraud column. They’re buried in headcount, overtime, and customer churn.

Why Traditional Defenses Fall Short

Retailers aren’t ignoring the problem – they just can’t keep up. Survey respondents average three or more prevention methods in place: employee training (58%), multi-factor authentication (49%), AI voice analysis (49%), and verification callbacks (49%). Yet 84% still faced sophisticated attacks in the past year. More layers aren’t translating to better protection.

Most of these tools were built for a less sophisticated threat environment. Caller ID verification doesn’t catch someone who’s already passed through it using a spoofed number. Knowledge-based authentication questions don’t help when fraudsters have harvested personal data from breaches. Both approaches slow down real customers while presenting only a minor speed bump to a motivated attacker. Deepfakes are making this challenge markedly harder: 74% of respondents have encountered deepfake or voice cloning incidents. 

“As fraudsters grow more sophisticated, better authentication isn’t necessarily going to stop attacks,” Pappas explains. “You need the ability to detect manipulation tactics being used on top of whatever authentication layer you have in place.” 

Henry LeGard, CEO of Verisoul, adds that new fraud patterns can go from isolated incident to large-scale operation within days, spreading rapidly through organized fraud networks—and that retailers need forensic analysis rather than relying on human or auditory intelligence alone.

Real-time conversational intelligence is what closes that gap: systems that analyze tone, urgency, stress patterns, and behavioral cues during a call, catching manipulation the moment it begins, invisible to legitimate customers.

The Path Forward

Budget trends tell their own story. Ninety-two percent of surveyed leaders are increasing voice fraud spending in the next 12 months, with 40% making significant investments in new detection technology. Leaders are prioritizing AI-powered voice detection (27%) and real-time capabilities (47%), tools that catch fraud as it’s happening rather than after the damage is done.

When evaluating solutions, Pappas advises looking for transparency: “Choose tools that tell you why they’re flagging a call, not just that they did. You need solutions that explain the risk, flag the moment, and let humans make smart calls fast.” 

Both Pappas and LeGard recommend partnering with specialists rather than building in-house. Retailers tackling this alone will lack the critical data and modeling capabilities that come from a vendor focused exclusively on voice-based fraud.

The practical starting point is an honest audit. Map current detection methods, identify the gaps, and quantify what fraud is actually costing in investigation hours, customer friction, and operational drag. The numbers are often more sobering than expected.

“Bad actors will always find creative ways in,” Pappas says. “You need a system that adapts as fast as they do. Without automated monitoring that detects patterns and flags risk in real time, you’ll always be reacting instead of preventing.”

In retail, margins are thin and loyalty is fragile. Waiting for fraud to happen – draining resources and repelling the customers you’ve worked hard to attract – is a fatal gamble.

Explore what’s possible

Learn how your peers are using real-time voice intelligence to shrink risk and boost user experiences—for employees and customers alike.

> Request a demo at Modulate.ai