Omilia raises $67M to scale its customer support platform
TechCrunch Ivan Mehta
Athens-based Omilia just landed $67M to grow its AI customer-support platform. It's betting on mixing simple tools with AI, not going all-in on chatbots like flashier rivals.
Everyone building AI for customer service seems to want to hand you a bazooka. Omilia would rather hand you a toolbox. The Athens-based company, which has been automating phone calls and support queries since 2002, just closed a $67 million Series B led by Expedition Growth Capital, and its pitch is refreshingly unglamorous: not every customer question needs a large language model thrown at it.
CEO Dimitris Vassos makes the case bluntly. Someone checking an account balance doesn't need generative AI reasoning through the request; they need a fast, cheap, reliable answer. Vassos contrasts this with companies like Sierra, Decagon and Parloa, which he says brand themselves purely as generative AI shops. His analogy: you might carry a bazooka, but if the enemy's standing right next to you, you need a knife instead. Omilia builds self-learning agents that blend older automation techniques with newer AI models depending on what the query actually calls for.
The funding math backs up the pitch. This is only Omilia's second raise, after a $20 million round from Grafton Capital back in 2020, and in the years since, annual recurring revenue has climbed tenfold to $60 million. Vassos credits that growth to unit economics that work for Omilia and its customers alike, which he argues is why the company hasn't needed to burn through outside cash the way some AI-support startups have. Clients on the roster include Capital One, Discover, RBC, DWP and PSEG.
Quick-service restaurants are the newest frontier. Omilia's voice-ordering tech now runs in more than 1,000 Taco Bell locations, and Vassos says two more U.S. chains are in active talks. That said, AI ordering systems have had their share of embarrassing moments industry-wide, including a widely reported case of a Taco Bell customer supposedly ordering 18,000 cups of water. Vassos flatly denies it happened on Omilia's system, pointing to logs that show no such order — though Taco Bell hasn't confirmed either way.
The new capital is earmarked for a U.S. office and a beefed-up go-to-market team, with Omilia hiring a chief revenue officer, chief marketing officer and VP of revenue operations. Headcount, currently around 500, is expected to hit 600 by year's end. Vassos isn't chasing LinkedIn buzz; he says the goal is steady growth toward a billion-dollar revenue company within three years.
My take
The bazooka-versus-knife line is the smartest thing anyone's said in the AI customer-service space in months, and it exposes a real problem with venture-funded chatbot startups: they're optimized for hype cycles, not for the boring 80% of support tickets that don't need a language model at all. Companies burning cash to make an LLM answer 'what's my balance' are solving a problem nobody actually has. Twenty years of legacy infrastructure turns out to be a genuine moat when everyone else is racing to look impressive on LinkedIn instead of profitable on a balance sheet.
Read more about this at: TechCrunch