The Situation

The contact center remains one of the most exploited channels for identity fraud. Knowledge-based authentication relies on information that is widely available through data breaches, social engineering, and dark web credential markets, giving agents no reliable way to confirm the person on the line is who they claim to be. Generative AI has compounded this exposure significantly. Synthetic voice tools allow fraudsters to impersonate account holders convincingly in real time. LLMs enable automated systems to conduct fluid, contextually aware conversations indistinguishable from human callers. AI-assembled fraud dossiers drawn from breached data, public records, and social media can be deployed at a scale and velocity no human fraud operation could match.

The fraud vectors active in the contact center include:

• Account takeover: bad actors using breached PII, AI-generated voices, and intercepted OTPs to convince agents or automated systems to make account changes, reset credentials, or authorize transactions
• Social engineering of the agent: fraudsters manipulating agents into bypassing security procedures through fabricated scenarios, urgency, or synthetic voice impersonation, achieving account access or transaction authorization without ever triggering a formal authentication challenge
• Information harvesting: extracting account details, PII, or coverage information from agents without taking over the account, providing the raw material for downstream fraud across other channels

Knowledge-based authentication is no longer a reliable control for contact center security. Voice biometric matching without anti-spoofing layers shares its own blind spot: it cannot confirm the voice on the line is human.

The Solution

Daon provides contact center authentication that confirms the identity of the caller through biometrics with built-in anti-spoofing. This closes the gap that knowledge-based authentication and unaided voice biometrics leave open.

Daon’s contact center authentication offers a layered approach to fraud prevention, beginning with analyzing fraud signals from the incoming connection, then adding synthetic audio detection from the time the call connects, and finally confirming the identity of the caller. Pre-answer risk intelligence analyzes carrier, line, and device signals such as SIM swap activity and ANI spoofing to flag high-risk calls before the caller ever speaks. Dedicated signal-level synthetic audio detection analyzes frequencies, phase relationships, and digital artifacts to identify AI-generated audio regardless of language or cloning model, achieving 99% accuracy on known models. Voice biometric authentication identifies the caller against their enrolled voiceprint, with active and passive modes available depending on the organization’s channel design. A built-in anti-spoofing stack covers replay detection, synthetic speech detection, pre-seen audio detection, and distorted voice detection. Voice biometric authentication integrates natively with Genesys, Cisco, Amazon Connect, Avaya, Five9, and NICE, deploying into existing contact center infrastructure without replacing it.

For organizations that leverage the broader Daon suite of solutions and employ Identity Continuity, identities verified at enrollment can be bound to the voiceprint through face authentication at the time of enrollment. Step-up authentication via a push notification to the organization’s app can trigger a face biometric confirmation or other authentication factor at any point in the interaction, to add assurance for high-value transactions or to minimize false positives. Because all authentication events resolve to the same identity record, signals from the contact center are visible across every other channel in the lifecycle.

Beyond fraud prevention, replacing knowledge-based authentication with biometrics removes friction for legitimate callers, reducing abandonment and churn. The same shift delivers direct cost benefits: shorter calls, higher IVR containment, and reduced fraud losses each contribute directly to lower operational costs.