Omilia Raises $67M to Expand Voice AI

CMSWire reports that Omilia describes its platform as proprietary agentic Voice AI for enterprise contact centers, with more than 200 deployments. The company asserted that one Tier 1 U.S. bank uses the platform for more than 1 million calls daily, and that it can handle more than 50,000 concurrent voice interactions for a single client at sub-second latency.

Its listed platform capabilities include self-learning agents, Lexis generative text-to-speech, and support for compliance standards including FedRAMP, PCI-DSS, SOC 2, HIPAA and GDPR. CMSWire further reports that Omilia claims sub-45 ms latency for its Lexis generative TTS model and a pricing model without token-cost pass-through.

Expedition Growth Capital founder and managing partner Oliver Thomas said in a statement that Omilia's "Self-Learning Agentic CX" was delivering a step change in call containment and operational improvement, with glass-box auditability and cost predictability that he described as structurally difficult for other vendors to match.

TechCrunch reports that CEO Dimitris Vassos described Omilia's approach as using multiple technologies rather than deploying large language models for every contact-center task. He told TechCrunch that routine requests, such as balance inquiries, do not necessarily require an LLM.

That technical distinction matters in contact centers, where a production system may combine deterministic workflows, speech recognition, intent classification, retrieval, generative models, text-to-speech, escalation paths and audit logging. Across comparable enterprise deployments, teams commonly evaluate those components on latency, containment, reliability, compliance controls and unit economics, not solely on conversational quality.

Omilia's financing arrives amid increased competition among vendors seeking to automate customer calls, chats and messages. TechCrunch identifies Sierra, Decagon and Parloa among the startups active in the category. For ML and platform teams, the reported scale and compliance claims make independently validating throughput, latency distributions, fallback behavior and human-handoff rates central to any enterprise evaluation.