Share with your CMO
Omilia is betting that voice AI built from scratch for regulated enterprises, not adapted from general-purpose LLMs, is the right architecture for high-stakes contact centers. The company raised $67 million in a Series B led by Expedition Growth Capital, reporting it grew annual recurring revenue past $60 million, more than 10x since its Series A, without additional equity. Customers include Capital One, RBC, and Taco Bell. The funding backs a first U.S. office opening in late 2026, with Omilia’s agentic voice platform already handling over one million calls per day for a single Tier 1 U.S. bank.
What this means for your business
The question this funding round actually poses for CX leaders isn’t whether Omilia is real, the ARR and customer roster settle that, it’s whether proprietary voice stacks can hold their ground as hyperscalers pour resources into general-purpose voice APIs. If you’re running a contact center in financial services, healthcare, or utilities where a wrong answer triggers a compliance event, the architecture decision your team makes in the next 18 months will be harder to reverse than most vendor swaps. Omilia’s pitch lands squarely on that exposure.
The “glass-box auditability” framing from Expedition’s managing partner is doing real work here, not just investor theater. Most agentic AI platforms, meaning systems that reason through multi-step customer interactions autonomously rather than following a fixed script, pass token costs through to customers and depend on third-party LLMs whose behavior can shift with a model update. Omilia’s zero token pass-through pricing and proprietary stack are structural arguments against that instability. The Forrester Wave Leader designation in April and the Gartner Visionary placement in July arrived within weeks of Lexis, its own text-to-speech model running at sub-45ms latency. That sequencing looks deliberate: analyst validation, then a product that makes the validation harder to dismiss.
The two Five9 veterans hired into revenue and growth roles in July signal where Omilia thinks the enterprise sales motion actually lives, inside the CCaaS replacement cycle, not alongside it. If your current contact center platform contract is up for renewal in 2026 or 2027, that’s the specific moment Omilia is targeting. The falsification condition for the whole thesis is simple: if Microsoft or Google ships a compliant, auditable voice stack at comparable latency before Omilia converts its U.S. foothold into a defensible installed base, the proprietary architecture advantage shrinks fast.
Concept deep-dive: Call containment
Call containment is the share of inbound customer calls resolved entirely by an automated system without routing to a human agent. It’s the primary efficiency metric in contact center AI, roughly analogous to a self-checkout rate in retail. Higher containment means lower cost per interaction and faster resolution, but only if the automation is accurate enough that customers don’t immediately request an agent anyway. Omilia’s 70% deflection rate claim, cited in the market context, is the number that determines whether this platform pays for itself.
Based on reporting from Omilia Secures $67M Series B to Expand U.S. Voice AI, originally published 2026-08-06 11:03:00.
