$7.5M Saved: How One Payer Used Passive Voice Authentication to Improve Member Experience and Stop AI Attacks

[!KEY]
item: 60-75 sec| saved per call on trusted profile matches
item: 60% | relative increase in authentication rates
item: $7.5M | operational costs saved per 10M calls
[/]

Summary

A major U.S. payer was facing a dual challenge: inefficient, friction-heavy authentication was frustrating members, while high volumes of automated bot traffic were clogging the contact center. By deploying Pindrop® Passport, the organization streamlined the experience for legitimate members to authenticate passively and self-serve instantly. The results: $7.5M operational savings per 10 million calls, 60-75 seconds saved per call, and 1% increase in IVR containment.¹ By automatically clearing the "noise" of verified members, the payer isolates unauthenticated traffic, where Pindrop® Pulse can detect sophisticated AI-driven threats that bypass traditional authentication methods with ease.

Challenge

Cumbersome authentication created operational drag, hindered member self service, and left the organization unable to distinguish legitimate members from automated bot traffic.

Solution

Pindrop® Passport replaced that process with passive, real-time voice authentication built directly into the IVR, enabling a better experience and more self service. Their new Pindrop® Pulse implementation adds another layer of security to detect synthetic and deepfake threats.

Legacy authentication creates a poor experience and isn’t secure

Traditional, multi-step authentication was a major source of friction for members. With only ~35% of callers successfully authenticating inside the IVR, most were forced to wait for a live agent and repeat the entire identity verification process manually.

Like most healthcare organizations, this major payer had leaned on knowledge-based authentication (KBA) for years: member ID, date of birth, and zip, checked one at a time at the top of every call. Every failed authentication attempt meant 45–75 seconds of agent time spent re-verifying identity before the real conversation could even start, costing the enterprise, and members, millions of hours annually. 

This operational friction was compounded by a broader security blind spot: the inability to distinguish between a verified member, a routine third-party bot, and a malicious AI impersonator. Without passive verification or the ability to identify artificial voices, the contact center treated all callers with the same inefficient, manual workflow. Automated traffic, ranging from legitimate provider vendors to sophisticated attackers, clogged lines and trapped human associates. 

This issue was particularly acute in Medicaid, where the same associates handled both member and provider traffic and automated callers could occupy an agent for 20 to 30 minutes at a time. A single agent fielding just 5 automated calls a day is 2.5 hours; that’s 12.5 hours a week or 50 hours a month, and that’s 50 hours they are not helping members in need.

Most importantly, the system was fundamentally insecure. With attackers able to bypass KBA challenges 50% of the time, this manual approach left the organization exposed. Agents were forced to manually triage risk on every call instead of focusing their expertise on high-value member support.

Using passive authentication to enhance experience and security

Passive authentication works in the background to accelerate the member journey. With Pindrop® Passport, this organization verifies callers in real-time against a voice profile, device, and behavioral signals, to authenticate trusted members within the IVR without manual intervention.

This process helps filter contact center traffic. Legitimate members are authenticated instantly, clearing the noise from the queue. This operational advantage allows the organization to focus security resources and human agents exclusively on managing unknown or potentially risky callers.

Pindrop offered a way to secure authentication in such a way that callers can now eliminate the member ID option. They come in, give us two quick pieces of information, and we authenticate them. It's been highly successful.— Director of Contact Center Technology

As part of the change, the team built the program around explicit member opt-in, which paid off with high adoption rates.

Everything we do from a self-service perspective is always gaining that member’s buy-in on what we're offering. We generally would take a conservative approach, asking members to opt in. We've had very few people come back into the system and opt out.— Director of Contact Center Technology

Impact: ~60% relative increase in authentication rates

Passive authentication transformed the contact center’s baseline efficiency. Authentication rates in the mid 30% range climbed to the mid 50-60% range across lines of business.

This operational shift was the primary enabler for self-service. By resolving identity upfront, the payer migrated routine tasks—such as ID-card ordering—entirely into the IVR, where 80–85% of those requests are now completed without agent intervention.

When we started this journey, we struggled with simple things like ordering ID cards just because authentication was a challenge. But now ID cards are a non-issue. Those are self-served. They don't even get to a desktop in most cases. And then it allowed us to open doors for payment; that's another thing that became easier with auto authentication.— Director of Contact Center Technology

The change reached the agent desktop, too. With authentication resolved before a call ever connected, agents stopped opening interactions by confirming who they were talking to and started them by actually helping, reducing the need for KBAs and saving 60–75 seconds of handle time (~$7.5M per 10 million calls).¹

The next steps: Using AI for continuous security and operational improvements

The identity question the team started with was simple: is this caller who they claim to be? Inside the organization, that question is already being reframed into something broader: who, or what, is actually interacting with us, and how should we handle that interaction? 

In addition to analyzing, “Is this caller the right human?” this organization is implementing Pindrop® Pulse, to answer the question, “Is this caller a human at all?” These two security layers enable the workflow to handle non-human callers separately and securely. This includes the use of AI agents to field legitimate AI callers, such as AI bots acting on behalf of providers, while human callers go through the automated authentication and self service process, with staff focusing on those that are risky or need help. 

By automating the "knowns" and intercepting the "risky," the organization is effectively re-architecting the contact center. Human associates no longer have to untangle queues call-by-call; instead, they are freed to focus on complex health concerns and high-touch situations. Combined, these workflow capabilities create a secure, efficient, and intelligent loop: Detect, Authenticate, and Automate.

Is your authentication process ready to support self-service at scale?

Voice is where member experience is won or lost. Passive authentication brings trust back into the voice channel, yielding significant operational gains: agents spend less time on manual process and more time on high-value member support. By establishing a trusted lane for members and layering on advanced AI detection for the unknown, this enterprise is improving speed and building a foundation that makes it increasingly difficult for bad actors to penetrate their defenses.