ValidSoft details cryptographic intent binding for VoiceMFA in agentic AI

ValidSoft unveiled technical details of a cryptographic intent binding approach meant to prove not just who a caller is, but what action they authorized. The company says the VoiceMFA system is designed for contact centers, voice payments and audit workflows as agentic AI expands account and transaction handling.

Why it matters: - Agentic AI can take actions on behalf of users, which raises the bar for authorization controls. - Existing identity checks can confirm who is speaking, but they do not cryptographically prove what the person approved. - ValidSoft’s approach is aimed at creating a verifiable record of human intent for transactions and AI-driven actions.

What happened: - ValidSoft announced technical details of Cryptographic Intent Binding for VoiceMFA™ (See-Say). - The announcement was made Aug. 18, 2026, in Hartford, Connecticut. - The architecture is designed to verify human intent within agentic AI environments. - Pat Carroll, ValidSoft CEO, said the system is built to close the gap between identity verification and authorization.

The details: - VoiceMFA™ (See-Say) uses cryptographic hashing and device-bound codes. - The system binds a verified voice identity to the content of a specific authorized action. - ValidSoft says each authorized transaction or instruction is designed to be genuinely authorized, provable, non-repudiable and immutable. - “Provable” means the record can be demonstrated to a third party, including a regulator, auditor or customer, after the fact. - “Non-repudiable” means the authorizing party cannot credibly deny the intent captured at the moment of authorization. - “Immutable” means the record cannot be altered, deleted or retrospectively modified. - VoiceMFA™ (See-Say) is designed to be deployed inside ValidSoft’s AI Voice Identity Platform, or VIP™. - VIP™ also includes Voice Verity® for deepfake and synthetic speech detection. - VIP™ also includes VoiceID™ for voice biometric identity verification. - Each capability can operate as a standalone solution. - ValidSoft says using VoiceMFA™ (See-Say) with Voice Verity® and VoiceID™ is designed to provide layered coverage across detection, identity and authorized intent. - The framework is intended for contact centers, voice and agentic payment channels and compliance audit workflows. - The system is designed to create a record of the identity and authority under which a transaction or AI agent action was carried out.

Between the lines: - The announcement signals a shift from identity assurance to authorization assurance. - That matters as AI agents increasingly handle actions that can create financial, compliance and customer-risk exposure. - The platform framing suggests ValidSoft is positioning voice security as a stack, not a single feature.

What's next: - ValidSoft is expected to push VoiceMFA™ (See-Say) as part of broader deployments of VIP™. - The main use cases to watch are contact center authentication, payment authorization and audit evidence collection. - Wider adoption will likely depend on whether the system can be integrated into existing voice and AI workflows without adding friction.

The bottom line: - ValidSoft is betting that proving intent, not just identity, will become a core security requirement for agentic AI.