VoxEQ - Voice Intelligence Solution - VoxEQ® logo

Credit union multi-layer fraud defense: webinar recap with TTEC Digital, TransUnion, IDgo, and VoxEQ

Introduction

This recap distills key takeaways from Jack Caven’s discussion with TTEC Digital, TransUnion, and IDgo on building layered defenses against member impersonation and account takeover in credit unions. The panel showcased how TTEC Digital’s SmartApps Cloud orchestrates best‑of‑breed controls—including VoxEQ Verify voice biometrics, TransUnion identity/device risk, and IDgo mobile MFA—to stop fraud while reducing friction for legitimate members. See the on‑demand session on BrightTALK for the full demo and Q&A. Watch the session.

Why credit unions need layered fraud defense

  • Fraud has shifted to the voice channel; attackers blend breached data, social engineering, and AI voices to bypass static checks. VoxEQ’s overview of the problem and approach is summarized in the Fraud Detection Playbook. Read the playbook.

  • Legacy voiceprint programs protect only the enrolled subset of members, leaving gaps. Modern voice biometrics should add enrollment‑free screening and anomaly detection for 100% of calls. What is voice biometrics?

  • Regulatory and risk teams increasingly expect two or more factors spanning knowledge, possession, and inherence—ideally orchestrated with minimal added handle time. ID/V vs. fraud detection

  • CU member experience matters: cutting seconds from ID&V, minimizing re‑authentication, and routing risky calls to specialists improves trust and CSAT. For CX guidance, see the Voice‑Led CX Playbook.

Roles of each layer demonstrated in Smart

Apps Cloud | Layer | Primary role in the stack | When it runs | What it uses | Outcome | |---|---|---|---|---| | Orchestration (TTEC Digital SmartApps Cloud) | Coordinate checks, route outcomes, minimize agent effort | Continuously across the flow | Connectors, rules, and UI for agents | Unified workflow that blends risk, CX, and step‑up actions. Partnership announcement | | Voice biometrics (VoxEQ Verify) | Screen every caller for profile–voice mismatch; detect impostors and synthetic voices; match against Watch List | First seconds of live audio | Physiological bio‑signals from voice; no PII/voiceprints stored | Real‑time risk signal, language‑agnostic, enrollment‑free. VoxEQ Verify · Genesys AppFoundry | | Identity/device risk (TransUnion) | Device, identity, and behavioral risk to corroborate or elevate voice risk | Parallel to call start or pre‑agent | Device and identity telemetry, risk models | Risk score that informs routing and step‑up (knowledge factor optional). | | Possession factor (IDgo mobile MFA) | Step‑up authentication when risk is elevated | On demand, triggered by policy | Member’s mobile authenticator | Cryptographic bind to member device to approve sensitive actions. |

Note: VoxEQ operates only after the call connects and the caller speaks; it does not rely on telecom metadata, STIR/SHAKEN, ANI, or phoneprinting. Its signals are derived from the sound of the live voice.

How VoxEQ “screens every caller” without enrollment

  • Enrollment‑free screening: Verify analyzes bio‑signals in the first seconds of audio to flag profile–voice mismatches (e.g., age/birth‑sex inconsistencies with the claimed member) for both known and first‑time callers. VoxEQ Verify

  • Privacy‑first: no storage of member PII or voiceprints; outputs are labels and risk scores. AI Ethics

  • Watch List: unattributed fraudster voices and synthetic signatures help catch repeat impostors in real time. Product guide

  • Deepfake/synthetic detection: continuously refined, while allowing legitimate synthetic use cases (e.g., voicemail). Verify overview

Sample CU flow (end‑to‑end)

1) Call connects → SmartApps begins orchestration. TTEC Digital + VoxEQ 2) 0–3 seconds of audio → VoxEQ Verify outputs risk signal; if profile–voice mismatch or synthetic is suspected, risk elevates. AppFoundry listing 3) Parallel device/identity risk (TransUnion) returns corroborating score. 4) Policy decision:

  • Low risk → streamline ID&V, proceed with normal servicing.

  • Medium risk → agent receives guidance; knowledge‑based element may be applied as needed. ID/V vs. fraud detection

  • High risk → step‑up via IDgo mobile MFA (possession factor) before any sensitive action. 5) Outcomes auto‑logged; Watch List updates if a fraud attempt is confirmed. Product guide

KPI checklist for CU leaders

Security and fraud

  • Account takeover attempts detected (volume, rate)

  • Confirmed impostor intercepts; Watch List hit rate

  • Synthetic/deepfake detection rate and false positive rate

  • Coverage: percent of inbound calls screened by inherence (target: 100% with VoxEQ Verify)

CX and operations

  • ID&V time saved (seconds per call) and impact on AHT

  • First‑call resolution (FCR) and containment for risky vs. non‑risky calls

  • Agent effort: steps/clicks removed by orchestration

  • Escalation rate after step‑up MFA

Compliance and governance

  • Authentication factor coverage across knowledge/possession/inherence

  • Privacy posture: no storage of PII/voiceprints by the biometric layer

  • Auditability: decision logs and policy traceability

Measurement notes

  • Instrument SmartApps to capture per‑call risk signals, policy decisions, and handle‑time segments.

  • Segment results by line of business (e.g., wires, card disputes) and by member cohort to tune policies.

Implementation notes (speed, privacy, platforms)

  • Speed to value: teams have deployed the combined solution rapidly—often within a day—using SmartApps Cloud and VoxEQ’s stable API. Deploy in a day

  • Platform fit: Verify, Persona, and Prompt are available via Genesys AppFoundry and integrate with leading CCaaS stacks.

  • Privacy by design: VoxEQ provides risk scores/labels only; member PII and content remain in your systems. AI Ethics

Watch, read, and take next steps