The Company That Made Me Stop Scrolling
I look at LinkedIn job postings the way people used to read newspaper classifieds.
Not because I'm looking for work. Because the job postings tell you something the press releases don't. Which companies are expanding. Which are quietly contracting. Which roles are being created that didn't exist two years ago. Which skills suddenly command a premium. The job board is a real-time index of where money is moving and what problems people are willing to pay to solve.
I've written before about my complicated relationship with LinkedIn — I hate the platform and I use it every day, which is a contradiction I've made my peace with. If you haven't read that piece, it's here. The short version: LinkedIn is built on performance and I find performance exhausting. But the job data is real, and the job data is useful, and so I keep scrolling.
On one of those scrolling sessions, a company stopped me.
The name stopped me first. Wonderful.ai. I care about domain names in a way that most people find excessive — a name that works as a domain, that tells you something about the product, that sounds right when you say it out loud, is genuinely hard to find and genuinely valuable when you do. Wonderful.ai is all of those things. It doesn't tell you what the company does but it tells you how they want you to feel about it. That is a different and sometimes better signal.
I looked them up. Then I talked it through with my AI agent. Then I decided this story needed to be written.
The Thesis Nobody Else Was Running
Wonderful was founded in early 2025 by Bar Winkler and Roey Lalazar — both Israeli, both based in Amsterdam, both with backgrounds in building enterprise software that gets acquired. Winkler previously built Approve.com, a platform for automated payment management, which was acquired by Tipalti in 2021. Lalazar co-founded Kaps, an AI language localization company. Between them: financial automation and language intelligence — exactly the two capabilities you need to build AI agents that work across cultures and markets. They knew what it looked like to build something enterprises will buy.
The company's founding thesis was specific. Most enterprise AI platforms are built in the United States, designed around US enterprise culture, US regulatory frameworks, and US enterprise buying patterns. When those platforms try to expand globally, they hit what Wonderful calls "AI deserts" — markets where the US-centric approach simply doesn't work. Different languages, obviously, but more than that: different cultural norms around how business conversations happen, different regulatory requirements around data and compliance, different expectations from enterprise buyers about how software gets implemented and supported.
Wonderful's answer was not to build a more flexible platform. It was to hire locally. Physical deployment teams in each market — people who understand the culture, the language, the regulatory environment, and the enterprise buying process from the inside. Not a global product with local configuration. A global platform with genuine local presence.
This is harder to build than a software platform. It is also harder to copy.
What Eight Months Looks Like
Wonderful emerged from stealth in mid-2025 with a $34 million seed round. Four months later, Index Ventures led a $100 million Series A. Four months after that, Insight Partners led a $150 million Series B at a $2 billion valuation.
$284 million in total. Eight months from seed to $2 billion valuation. 30+ countries. 650 employees, scaling to approximately 900 by year end.
The speed is not the interesting part. The interesting part is that the growth is happening in markets that other enterprise AI companies have found difficult: telecom in the Middle East, financial services in Southern Europe, manufacturing in Central and Eastern Europe. These are not easy markets for a new entrant. Wonderful is growing in them because it showed up with people who understand them, not just software that was designed somewhere else.
In February 2026, Wonderful partnered with Deepdub — a foundational voice AI company — to integrate voice synthesis directly into its multilingual agent framework. The move extended Wonderful beyond chat and email into autonomous call center workflows. Enterprise AI that can handle a phone call in Dutch, Arabic, or Portuguese, in the specific register a local customer expects, is a different product from enterprise AI that handles a chat window.
The retention number that tells you whether the thesis is working: over 70% of enterprise customers expanded to additional operational workflows within three months of onboarding. They came for customer support. They stayed for everything else. That is the horizontal platform play working exactly as designed.
Bar Winkler put it plainly: "The promise of AI agents is clear, but putting that into practice — and critically, into production — is a huge challenge. It requires marrying best-in-class technology together with flawless delivery, on the ground with customers."
On the ground. That phrase is doing a lot of work in their story.
What They Actually Build
Wonderful builds enterprise AI agents — systems that handle customer interactions across voice, chat, and email, integrated with the enterprise's existing systems, able to update records, resolve issues, escalate when necessary, and operate within the governance and compliance frameworks the enterprise already runs.
The customer support use case was the entry point. The platform is expanding into employee training, regulatory compliance, sales enablement, and onboarding. Each new use case runs on the same foundational infrastructure — the local deployment teams, the cultural calibration, the compliance integration — which means each new use case costs less to add than it cost to build the first one.
This is the horizontal platform play. Build the foundation once. Stack use cases on top. The more use cases a customer deploys, the more embedded the platform becomes. The more embedded, the stickier the relationship. The stickier the relationship, the harder the comparison to a competitor who hasn't built the same local presence.
The job posting that caught my eye was for Malta. Forwarded Deployment Manager. That title tells you something about the model — not a remote account manager, not a sales representative, but someone embedded locally to manage the deployment of agents in a specific market. Malta is small, but it is a financial services hub, a gaming technology hub, and a company formation hub for European operations. The enterprises Wonderful wants to serve are here.
Why This Story Needed to Be Written
I have a rule about the companies I write about in Off The Record. The story has to have something in it that changes how you think about the category — not just what the company does, but why the way they do it matters.
Wonderful's thing is the local embedding. The belief that AI deployment is not a software problem but a people problem, and that the people problem requires physical presence, not configuration options. This is a bet against the dominant model in enterprise software — the one that says you build a flexible global platform and let customers adapt to it — and a bet in favor of something that looks more like professional services than SaaS.
Professional services don't scale like SaaS. Wonderful is betting that the platform underneath the service does, and that the local presence is the moat rather than the cost center. If they're right, the model is very hard to replicate. If they're wrong, $284 million is a lot of money to spend on deployment teams.
Eight months in, the market is suggesting they're right.
I started FreeMalta eight months ago. Wonderful exited stealth eight months ago. I note this not because it means anything in particular, but because the timeline is the same and the lesson is not. Eight months is enough time to go from nothing to $2 billion if you have the right thesis and the right execution. It is also enough time to build something meaningful from one person, one server, and a clear idea of what the gap in the market actually is. The scale is different. The principle is the same: find the thing that isn't being done well, do it properly, and don't stop.
That is the story worth paying attention to. Not the funding round — funding rounds are common enough to be noise. The thesis, and whether the evidence is supporting it.
It is. So far.
FreeMalta covers company founding stories at The Garage. Wonderful's full story is there. For enterprise AI deployment and digital transformation services in Malta, visit maltainsider.com.
