Ex-Google engineers raise $8M for Klarent to automate enterprise software testing
Tech Funding News Sofia Chesnokova
Ex-Google engineers raised $8M for Klarent to automate software testing. It uses AI to write tests, not run them, and says that scales where outsourced QA doesn't.
Based on reporting by Tech Funding News, Sofia Chesnokova — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Klarent has pulled in $8 million in seed funding to go after a job most teams still treat as grind work: software testing. Mosaic Ventures led the round, with Moonfire Ventures also in, and the Zurich startup says it wants to chip away at outsourced quality assurance by turning tests into something AI can generate and reuse.
The company’s twist is pretty specific. Klarent says its agents write test code, cache what they learn, and then reuse it, rather than making model calls while software is running. That’s a clean split from what founder Asheem Panakkat calls a probabilistic approach. In Klarent’s view, that style may look clever in a demo, but it falls apart when you try to use it at enterprise scale.
Panakkat knows the enterprise side from the inside. He spent years at Google as a software engineer and engineering manager, working on Maps, Search, Shopping and Lens. In 2023, he and colleague Momchil Ivanov left to start Klarent. Their first idea was evaluation tools for large language models, but they decided the bigger opening was testing the code those models generate.
The market they’re targeting is large and old-school. Panakkat says quality assurance is a $35 billion business, much of it still manual. Mosaic partner Chandar Lal went further, arguing that outsourced testing is part of a much bigger services market that relies on people clicking through huge numbers of scripts and buttons. Klarent wants to replace that with what it calls “deterministic autonomy,” while keeping a human-in-the-loop service checking results.
The company already has a customer doing about 3,500 tests a day, and Panakkat says the system can follow workflows from placing an order on a website through order management and delivery systems. Klarent has 25 employees today, and most of the new money is earmarked for hiring in the United States, with sales expansion in the US and UK, plus some engineering. The angel list includes Harris Barton, AngelInvest Ventures and Jürgen Galler, described in the announcement as a former Google director.
And Klarent is making a pretty bold bet: that testing will stop being a cost centre and become the thing that decides whether software can be trusted at all. That’s a good pitch if you believe AI will keep writing more code. It’s also a familiar startup move: take a giant, miserable enterprise task, call it inevitable, and hope the spreadsheet agrees.
My take — AI-written commentary, not fact-checked reporting
This is the kind of AI startup pitch that actually makes sense: less magic, more boring infrastructure. Writing tests is a better use of models than letting them freewheel through production and hoping for the best, which is how a lot of “agentic” software ends up sounding brave and looking expensive. Outsourced QA is a big target, and that alone makes Klarent more credible than the usual demo-driven fluff.
Read more about this at: Tech Funding News