The news
Athens-based Omilia, which has been automating voice calls and customer support since 2002, has raised a $67 million Series B led by Expedition Growth Capital. It is the company's second major raise: a $20 million round from Grafton Capital in 2020 was followed by annual recurring revenue growing 10x to $60 million.
The customer list reads like a who's who of enterprise support: Capital One, Discover, RBC, DWP, and PSEG, plus Taco Bell, where Omilia handles voice ordering across more than 1,000 outlets. Talks are underway with two more US quick-service restaurant chains. The new money will fund a US office, a go-to-market team, and senior hires including a CRO, CMO, and VP of revenue operations, with headcount growing from about 500 to 600 by the end of the year.
Why it matters
"You may have a bazooka, but if your enemy is near you, you need a knife. This is the reality of the contact center, where you need multiple tools." — Dimitris Vassos, Omilia CEO
Vassos's point is a direct challenge to the throw-an-LLM-at-everything school of AI. Omilia's data shows a large proportion of support queries are basic information requests, like account balances, that never needed a large language model in the first place. The company's bet is that durable automation comes from matching the cheapest reliable tool to each job — rules and speech recognition for simple queries, LLMs only where real judgment is required.
That stance stands out in a market full of startups bolting generative AI onto every support ticket. Omilia has spent more than two decades building speech recognition and conversational infrastructure, and it treats LLMs as one more tool in that stack rather than the whole stack.
That philosophy is showing up in the numbers: 10x ARR growth since 2020 and enterprise clients across banking, utilities, government, and restaurants.
What it means for Pakistan
Pakistani businesses wrestle with the same problem Omilia solves, just at a smaller scale: repetitive calls asking about order status, balances, timings, and FAQs. The lesson from this round is to right-size automation. A simple AI phone agent can handle routine queries, and AI automation services can route the rest — without paying premium model prices for every interaction.
The same right-sizing logic applies to a Karachi retail store as to a multinational bank: measure what callers actually ask, automate the top repeat queries first, and free your staff for the conversations that build trust.
Creators can borrow the multi-tool mindset too: ElevenLabs (PKR 3,300/mo at AI Tools Pak) is excellent for realistic voice, but you don't need it for every notification. If you want help scoping a phone agent or WhatsApp automation for your business, message AI Tools Pak on WhatsApp (+92 371 454 9245 or wa.me/923714549245) for current setup pricing.
Clients and use cases
Omilia's client roster spans banking, utilities, government, and quick-service restaurants:
| Client | Sector / use case |
|---|---|
| Capital One | Banking customer support automation |
| Discover | Banking customer support automation |
| RBC | Banking customer support automation |
| DWP | UK government support |
| PSEG | US utility customer service |
| Taco Bell | Voice ordering across 1,000+ outlets |
Practical takeaway
A simple way to copy Omilia's approach in a small business:
- Audit your inbound queries before buying AI: count how many are simple information requests.
- Start with the cheapest reliable tool for each query type, not the most powerful model.
- Deploy a phone agent for routine calls and keep humans for edge cases.
- Scale spend only after you measure deflection and resolution rates.
None of this requires a six-figure budget. The tools that handle routine queries are cheaper than ever, and the savings show up where Omilia's clients see them: lower cost per resolved ticket.
DEEPER DIVE
- TechCrunch on Omilia's $67M raise
- AI Tools Pak — AI phone agent
- AI Tools Pak — AI automation services
- How to automate WhatsApp with AI in Pakistan
Frequently asked questions
How much did Omilia raise and who led the round?
Omilia raised a $67 million Series B led by Expedition Growth Capital, following a $20 million round from Grafton Capital in 2020.
Why does Omilia avoid using LLMs for every query?
Because most support queries are simple information requests like account balances that do not need a large language model. Omilia matches the cheapest reliable tool to each job and uses LLMs only where judgment is needed.
Which big clients use Omilia?
Clients include Capital One, Discover, RBC, DWP, and PSEG, plus Taco Bell, which uses Omilia for voice ordering across more than 1,000 outlets.