Zenity

From working together at Microsoft to a company of their own

Ben Kliger and Michael Bargury did not meet in order to found Zenity: they had already worked together at Microsoft, on security products for the cloud and for operational technology (OT). That shared background matters for understanding the timing: the two saw from the inside how large organizations try to secure complex environments, and built on that experience when they founded Zenity in 2021, a year when generative AI tools had barely begun to enter wide use in organizations. In other words, the company was founded before the problem it solves became mainstream, a bet that paid off when the wave of AI agents arrived at large scale in 2024–2026.

Not blocking a model, but supervising an action

The difference between securing a language model and securing an AI agent is fundamental. A language model answers questions; an AI agent actually takes actions: it sends an email, changes a file, accesses a database, runs a business process. Zenity does not try to censor what the model says, but to supervise what the agent does: to understand the intent behind each requested action, and to decide deterministically whether to approve it, modify it or block it, before it is actually carried out, not after. This approach differs from traditional security tools that rely on known threat signatures, because it has to deal with actions an AI agent can invent in real time and that have never been seen before.

A $125 million round, and about $185 million in total

Zenity had raised smaller amounts before, but the August 2026 Series C, $125 million led by Norwest, was the largest in its history and brings its total funding to about $185 million. The list of investors in this round is interesting in itself: alongside classic venture capital funds such as Vertex Ventures and Third Point Ventures, the investment arms of industrial and technology corporations also joined: SoftBank Vision Fund 2, Hitachi Ventures and LG Technology Ventures. Such a mix of strategic, not just financial, investors usually indicates that the company is already selling to customers in industries where these corporations have a direct presence.

Customers among the world's largest companies

According to the company's announcement, Zenity's platform is used by Fortune 500 and Global 2000 organizations in finance, healthcare, pharmaceuticals, manufacturing and technology. Such figures are hard to verify independently (they come from a company announcement, not an audited report), but the size of the round and the identity of the strategic investors are consistent with a claim of presence at large organizations. With about 230 employees worldwide, about 150 of them in Israel, the company is still small relative to the size of the customers it says it serves, a ratio typical of enterprise software companies at an early growth stage.

"The company to beat" in a new field

Gartner analysts described Zenity as "the Company to Beat" in AI agent governance, wording that points to an early lead in a market that is still very young. It is important to understand what such a description means: it does not say the product is technically the best, but that the company managed to define the category in the eyes of analysts and potential customers before competitors could do so. In a young market category like "AI agent governance," that position is worth a lot, but it is also vulnerable, because a new category quickly attracts both large competitors and existing security vendors that expand their products.