The telephony platform duplicates the call audio stream to the FARx SIPREC Recording Server. No call quality is affected. An integrated telephone system can support caller verification by using Automatic Number Identification (ANI) so the telephone system pulls the customer record before the live conversation. Supported natively by Cisco UCM, Genesys, Avaya, Ribbon, Oracle ACME, AudioCodes, and most SIP-compliant platforms. These integrations can also work with IVR flows that use automated checks to match incoming numbers against stored customer records and collect identifying information before a caller reaches an agent.
The recommended WebSocket path. The FARx Broker handles session lifecycle, audio buffering, and intelligent chunking, removing all audio processing complexity from the integrating application. Compatible with Twilio, Amazon Connect, Five9, NICE CXone, Talkdesk, and Vonage.
For teams that prefer full control over the audio pipeline. The application handles chunking and posts segments directly to the FARx REST API. Well-suited to bespoke implementations with existing audio processing logic.
Every audio frame is evaluated simultaneously across all five dimensions and fused them into a single, continuously updated trust score - returned via REST API, WebSocket, or webhook. AI can also support the wider security process by analysing live conversations for sentiment and compliance signals.
Live voice matched against the enrolled voiceprint for that customer or employee - hundreds of unique acoustic features that cannot be replicated by mimicry or script-reading. Detected within the first few seconds of speech.
Fraudsters who correctly answer security questions are still caught. FARx verifies identity independently of what the caller says - correct PINs or passwords cannot by themselves prove the caller's identity or override a biometric mismatch.
AI-generated speech identified through acoustic analysis of spectral inconsistencies, vocoder artefacts, and generation model fingerprints is inaudible to the human ear but reliably detectable at the signal level. The detection technology is continually updated and retrained to recognise the latest AI-generated voices as new voice-cloning and deepfake tools enter the market. Organisations should incorporate voice biometrics with controls that protect against deepfakes and replay-based spoofing as part of robust ID verification.
Pre-recorded genuine voice samples distinguished from live speech through temporal inconsistency analysis, channel characteristic mismatches, and the absence of natural conversational micro-variations.
Each inbound call is screened against voiceprints of known bad actors in parallel with identity verification. A match triggers an immediate alert before any account interaction proceeds. Flagged matches can automatically flag suspicious calls for review or escalate suspicious calls before any sensitive action is approved.
FARx for Telephony integrates with virtually any telephony environment already in use, covering the full range of enterprise and cloud platforms via SIPREC and WebSocket.
Verifies the caller's voice silently during the Interactive Voice Response (IVR) interaction, before reaching a human agent, where interactive voice response can use keypad or speech recognition to let callers input identifying information and perform initial checks. This means fraudsters who pass security questions are still caught before accessing accounts or initiating transactions. Faster caller authentication in customer calls can cut call handling time, with caller verification reducing fraud losses by an average of 40% in high-risk environments.
Replaces or augments PIN and KBA challenges in automated IVR flows; while one-time passwords (OTPs) are sent to registered mobile numbers for verification, they can be intercepted or bypassed through SIM swap attacks, so other verification methods are often needed. Genuine callers move through faster, and once the customer passes verification, they benefit from the same process across channels rather than repeating checks; fraudsters are blocked before reaching sensitive data or high-value transaction options.
Passive voice verification throughout the call ensures the caller is the registered patient or authorised representative - without requiring them to remember PINs or answer personal questions. With multi-factor authentication layered in for higher-risk scenarios this helps verify identity in regulated healthcare workflows while supporting patient safety and reducing fraud exposure; in practice, multi-factor authentication reduces fraud losses by an average of 40%.Anyone sitting adjacent to the user in a café, airport, or train triggers automatic screen blur - no privacy screen required.
Meets identity assurance requirements under UK Gov.UK One Login, NIST SP 800-63, and eIDAS for benefits, tax, licensing, and social care telephone services, without requiring digital channel migration.
The same process can carry across phone or digital channels to support a joined-up customer journey for public services.
Protects customer service calls where identity verification is required before SIM swaps, number porting, or account changes, including sensitive access requests and high-risk updates to contact details. Also, screening for AI-generated voice clones used to bypass existing controls.
The same process can carry across phone or digital channels to support a joined-up customer journey for public services.
FARx for Telephony is packaged as a Docker container, deployable across cloud, on-premises, private cloud, and hybrid environments, ensuring voice biometric data stays within your own perimeter when required.
FARx managed. Fastest path to production. Suitable where strict on-premises requirements do not apply.
Full deployment within your own data centre. No voice data, templates, or audit logs leave your perimeter.
Dedicated tenancy on AWS, Azure, or GCP. Cloud scalability with biometric data sovereignty.
Core processing on-premises with cloud burst capacity for contact centres with variable call volumes.
Most organisations complete a working proof of concept against their telephony platform within one week of receiving sandbox credentials.
Explore the FARx API