Omilia’s $67M Series B: Voice-First Agentic CX Without Token Bill Shock

Omilia raised a $67 million Series B led by Expedition Growth Capital on August 6, 2026. The Athens/New York voice-first agentic CX company says live ARR topped $60M — >10× since Series A — with capital efficiency rare in the GenAI CX arms race.

Key facts

Company

Omilia — Self-Learning Agentic CX

Round

$67M Series B

Date

August 6, 2026

Lead

Expedition Growth Capital (London / Boston)

Prior

~$20M Series A (Grafton Capital, 2020)

ARR

>$60M live (company)

Scale claims

Tier-1 U.S. bank >1M calls/day; insurer agent-assist >600k calls/day

Recognition

Forrester Wave Leader (Conversational AI for Customer Service, Q2 2026)

Who uses the product — and for what job

Users: contact-center and CX leaders at banks, insurers, utilities, government, and QSR.

Job: contain and resolve high-volume voice (and omnichannel) interactions with sub-second latency, auditability, and predictable cost — not a pilot chatbot that explodes the LLM bill.

Named logos include Capital One, Discover, RBC, DWP, PSEG, and Taco Bell (1,000+ drive-thrus / 38 states per company).

Omilia’s wedge vs “LLM for every utterance”: use the right tool for the job — deterministic flows for balance checks, generative agents where judgment is needed — on owned models so token pass-through does not own the P&L.

Why now

• Enterprises graduated from GenAI demos to production call volume.
• Token-priced CX stacks create CFO pushback at million-call scale.
• Regulated buyers want FedRAMP / PCI / HIPAA / GDPR paths, not lab sandboxes.
• QSR voice ordering and bank containment are measurable ROI stories, not vanity NPS.

Why Expedition — portfolio fit (growth equity lens)

Expedition markets itself as a transatlantic software & AI growth specialist ($1B+ AUM) that partners with founder-owned, capital-efficient leaders. Omilia’s six-year gap from Series A to a $67M B with $60M+ ARR is the textbook “use capital to accelerate, not survive” profile.

Likely founder rationale: pick growth equity that understands enterprise GTM and founder control, not a mega-fund that would force a burn race against Sierra/Decagon branding.

Competitive map

CEO Dimitris Vassos’s framing (via TechCrunch): you may have a bazooka, but contact centers still need a knife.

Market signal

~$60M ARR → $67M raise implies growth equity pricing discipline, not 2021 AI multiple theater. The category winners may be the vendors that own unit economics at 1M+ calls/day.

When not to treat Omilia as the only template

• Wrong for greenfield consumer chatbots with no concurrency or compliance burden.
• Wrong if your only proof is a LinkedIn demo — Omilia’s buyers care about FedRAMP and call containment.
• Wrong to ignore QSR edge cases; voice ordering still needs human escalation design.

Practical takeaway

• Founders (CX AI): Sell predictable cost + audit trails; owned models can be a CFO feature.
• Investors: Prefer capital-efficient voice platforms with named regulated logos over narrative-only agent demos.
• Operators: Benchmark containment and cost-per-resolved-contact at your peak concurrency, not vendor showcase traffic.