Omilia releases Lexis, a generative text-to-speech model built natively into its Cloud Platform — not via third-party API — delivering sub-45ms latency, 25+ languages, instant voice cloning, and PCI-DSS, HIPAA, and GDPR compliance for regulated enterprise contact centers.
On July 8, 2026, Omilia — a conversational AI provider for enterprise contact centers — launched Lexis, a generative text-to-speech model built natively into the Omilia Cloud Platform (OCP). Unlike most enterprise TTS deployments that route audio through third-party APIs, Lexis operates entirely within OCP's perimeter, giving organizations full sovereignty over their voice data, brand identity, and performance guarantees. The model is available immediately to existing OCP customers and included in new platform deployments.
Lexis delivers sub-45-millisecond first-audio latency for real-time conversational flow, supports more than 25 languages and locales with native pronunciation and prosody, and achieves under 3% word error rate on transcription. The model's context-aware architecture reads tone and emotion at the sentence level, adjusting pacing, intonation, and emphasis in real time rather than generating flat, uniform speech. A custom voice cloning feature creates a deployable branded voice from a short recording in minutes — a process that traditionally takes months. Omilia also introduced CX Ambassador Personas, a library of pre-tuned voices optimized for specific contact center roles such as 'The Trusted Advisor,' 'The Bright Concierge,' and 'The Quick Operator.'
The compliance architecture is the key differentiator. Because Lexis runs natively within OCP, audio never leaves the platform perimeter — no third-party speech, language, or voice services are involved — eliminating the data leakage risk inherent in API-based TTS integrations. Lexis inherits OCP's full certification stack: PCI-DSS, SOC 2 Type II, ISO 27001, HIPAA, GDPR, and CCPA, making it deployable in regulated industries where sending customer audio to an external TTS provider would violate data residency or privacy requirements. Deployment options include fully managed SaaS and on-premises, with SLA-backed voice availability not subject to third-party model deprecation or outages.
CEO and co-founder Dimitris Vassos framed the launch as a data sovereignty play: 'Sovereignty, latency, and brand control can't be delivered by bolting on third-party APIs.' CPO Claudio Rodriques added that Lexis 'eliminates the latency of external round-trips, the compliance risk of audio leaving your perimeter, and the fragility of unowned voice libraries.' Omilia's enterprise customer base — including Capital One, Discover, Taco Bell, RBC, and the UK's Department for Work and Pensions — provides an immediate deployment surface. The launch signals a broader industry shift toward platform-native voice AI, where TTS is a built-in platform capability rather than a separate service to integrate.
