Pomegra
Zurich-based Klarent closed a €7.13M seed round led by Mosaic Ventures to scale its agentic software-testing platform and open a US presence.
Key Takeaways
- Klarent raised a €7.13M (about $8M) seed round, announced 9 October 2026.
- Mosaic Ventures led; Moonfire Ventures and AngelInvest Ventures also invested.
- Valuation, team size and revenue were not disclosed.
Lead
Klarent, a Zurich startup formerly known as fore ai, announced a €7.13M seed round on 9 October 2026. Mosaic Ventures, the London-based fund, led the financing. The company will put the money toward entering the United States, hiring engineers and extending its product across web and mobile applications.
The company was founded in 2023 by two former Google engineers. Asheem Panakkat, the CEO, led teams on Google Shopping and Lens. Momchil Ivanov, the CTO, worked as a staff software engineer focused on information retrieval.
Who Backed Klarent and at What Terms?
Mosaic Ventures led the round, with Moonfire Ventures and AngelInvest Ventures joining. Individual backers include Jürgen Galler, a former Google director, and Harris Barton. Klarent did not disclose valuation, equity sold or any earlier funding, so the round cannot be read against a prior mark.
The sum is modest by the standards of AI infrastructure seeds in 2026, but it is sized for a company with paying enterprise customers rather than a prototype. Klarent counts retailer JD Sports, Swiss media group NZZ and mobility company Sixt among its users. It also works with banks and insurers, a sector it plans to push further into.
What Does Klarent Actually Sell?
Klarent sells an autonomous quality-assurance platform in which AI agents write software tests that then run as ordinary code. The company says its agents use large language models only to author test code. Once written, the tests execute in a deterministic, repeatable way, which addresses the flakiness that makes model-driven testing hard to trust in production pipelines.
A human-in-the-loop verification layer sits on top, and customers pay only for verified outcomes. The platform also covers test management, test data management and CI/CD integrations. It tests both web and mobile applications, and it competes with established automation tooling and with newer AI-native testing startups.
Why Is Testing Drawing Investor Money Now?
Investors are betting that AI-assisted coding is moving the bottleneck from writing software to verifying it. Chandar Lal of Mosaic Ventures put it this way: "With AI, code generation is scaling faster than our ability to test it, so QA will become the key bottleneck."
Panakkat framed the founding thesis similarly, saying LLMs were improving quickly and the volume of software being created was about to expand sharply. That argument is plausible, but it remains a forecast. Test volume grows only if engineering teams ship more code, and buyers have a long record of treating QA as a cost line to be squeezed.
What Are the Risks for Klarent?
The main risk is differentiation. Generating test code with LLMs is a capability that coding assistants and incumbent QA vendors can add, and several well-funded startups chase the same pitch. Klarent's answer is determinism and outcome-based pricing, which shifts risk to the vendor and puts pressure on margins if human review remains heavy.
US expansion adds a second test. Sales cycles, security reviews and compliance demands for American enterprises differ from the European accounts that built Klarent's early references. The company has not disclosed headcount, revenue or customer count, so progress will be hard to measure from outside.
What Comes Next?
Expect Klarent to hire in engineering and open a US go-to-market operation over the coming quarters. Financial services is the stated target for deeper expansion, a sector where audit trails and repeatable test results carry real weight. A Series A would likely hinge on US logo wins and evidence that human verification costs fall as the agents improve.
Outlook
Klarent enters the US with €7.13M, a tier-one European lead and three recognisable enterprise customers. The thesis, that testing becomes the constraint as AI writes more code, is credible but unproven at scale. The next milestones are US customers, margin data under outcome-based pricing and a follow-on round priced against that evidence.
