Yair Kuznitsov once set a goal with an unusual unit of measurement. He wanted to give developers, compliance leaders and their colleagues a million hours back. No grand theory of the future was required to understand the appeal. Someone had to stop asking engineers to take screenshots.
He and Roi Amior had encountered the problem while working at IntSights. They were building a SaaS product, then helping establish that it met SOC 2 requirements. The software belonged to the cloud era. The work of proving things about it involved manual evidence collection. For people paid to build systems that run without constant attention, this was an awkward arrangement.
That irritation became part of the founding logic of Anecdotes, where Kuznitsov is co-founder and CEO. It also offers a useful introduction to him. He is interested in the distance between what technology can do and the repetitive work people still perform around it. His career has gradually made that distance a business problem, a leadership problem and, lately, an AI problem.
The engineer who needed a wider view
Kuznitsov studied computer science and economics at Bar-Ilan University. His early career ran through security research and research-and-development leadership in the Israel Defense Forces. He later managed an R&D group at IntSights, from September 2018 to October 2019. The sequence gave him years of experience building technical systems before he had to build a company around one.
Working in innovation broadened the job. Sales, product and marketing became part of his field of view. He has described this period as the point when he could finally look beyond an engineering department and understand more of how a business operated. Entrepreneurship required learning the neighboring disciplines, including the ones without a compiler to tell you where you had gone wrong.
His eventual founding relationships had history. He knew Eitan Adler from their time as group leaders in Israeli intelligence; he worked closely with Amior at IntSights. Adler joined the original founding team. These were people who had already seen one another work on difficult technical problems, rather than strangers discovering their compatibility after ordering matching company shirts.
The screenshot has an expiry date
The commercial case grew beyond the founders’ dislike of administrative chores. Evidence about a software system can become stale while a team is still assembling it. A setting changes. An employee leaves. An access permission survives longer than intended. A picture of yesterday’s configuration can be perfectly legible and already unhelpful.
Kuznitsov’s answer was to collect evidence from the systems themselves and organize it into usable data. The company emerged from stealth in February 2021. Early users included Lemonade and AppsFlyer. The premise was practical: give compliance teams a continuing view of what is happening, with less dependence on repeated requests to engineers and other colleagues.
A conceptual view of the approach, not a performance claim.
In his public writing, he also pushed for compliance to have a voice in business decisions. A March 2022 essay argued that expansion plans should include the people who understand the requirements of entering a market. Otherwise a company can approve an attractive opportunity before understanding the work needed to pursue it.
It was an interesting argument for someone whose original frustration involved bureaucracy. He was asking for compliance professionals to join the important conversations earlier. The work could support growth if the business planned for it. Leaving it until late in the process made everybody’s job more difficult, including the people who had already promised the launch date.
Someone three years further along
The founder’s own education continued outside the product. His advice to aspiring entrepreneurs has been to find a trusted mentor who occupied a similar position roughly three years earlier. The attraction is specificity: someone who remembers the immediate problems and can speak honestly about them. General encouragement has its uses; a person who recognizes the next difficult decision has others.
“You have to assume that you don’t know everything.”
Yair Kuznitsov
That admission accompanies an ambitious view of what can be learned. He describes entrepreneurship as a set of skills people develop, and has spoken about learning through both giving and receiving advice. The technical specialist has to become comfortable with a larger syllabus. Being accomplished in one part of the business does not excuse incuriosity about the rest.
By 2022, his writing about compliance leadership was addressing soft skills alongside technical competence. The subjects fit his own transition: understand stakeholders, communicate across functions, and keep sight of the business while attending to details. A system may be well designed and still struggle if the people expected to use it cannot see how it helps them.
His later reflections retain that emphasis on learning. In a post marking five years of building Anecdotes, he described curiosity as central to moving from a deeply technical background into business leadership. He also began with family, crediting their support and observing that his children barely remembered his life before the company. The calendar of a startup and the calendar of a household had been advancing together.
A piano, a stopwatch, another attempt
There are other places to find Kuznitsov’s appetite for making things. He plays piano and has said he would choose music if he could take another profession. His published biography also lists skiing among his interests. Code is only one outlet for the builder; a keyboard can mean more than one thing.
Running supplies a more numerical account. In a personal passage on the Anecdotes blog, he describes completing the London Marathon in 2021 in 3:59:41. He returned in 2024 after sixteen weeks of structured preparation and finished in 3:11:48. He had wanted to break three hours. He missed that target and recorded a substantially faster time.
He used the training app Runna, which predicted 3:12:30 after adjustments during preparation. His finish was forty-two seconds faster than the prediction. He turned the experience into an analogy for compliance: useful data, collected throughout a process, can help someone understand where they stand and what needs attention.
The appealing detail is the unfinished goal. A substantial improvement and a missed target can occupy the same result. Kuznitsov reported both. The stopwatch allows less room for celebratory editing than a funding announcement, and the numbers give his ambition something unusually concrete to argue with next time.
What the funding number actually means
In April 2025, Anecdotes announced an additional $30 million led by DTCP. It extended a $25 million Series B, bringing that round to $55 million and disclosed funding to $85 million. The three figures describe different things. Startup finance would be easier to follow if every announcement came with a small arithmetic lesson.
Initial Series B + extension = combined Series B.
$85m in total funding disclosed at that point.
There was a less decorative detail behind the extension. DTCP had made a smaller earlier investment and set milestones for the following year. Kuznitsov said the company met those conditions before the investor increased its commitment. The sequence linked the new capital to work already done.
His April 2025 Founder Storiez conversation covered the decision to extend the Series B, rather than move immediately to a Series C. It also returned to his development from technical expert to executive and the importance of trust in business partnerships. Financing changes the resources available to a founder; it leaves plenty of decisions that money cannot make on his behalf.
The next bottleneck has a person waiting beside it
He had been considering the role of generative AI in this work before the latest agent discussion. In June 2023, he published an article about using it in GRC, discussing potential applications around regulations, policies and risks, with safeguards in place. That earlier interest puts the current strategy in context. The question has developed from where generative tools might help into how software could carry a process forward. The recurring challenge is earning enough confidence for organizations to use the result.
By January 2026, Kuznitsov was describing a further obstacle. Collecting evidence automatically left substantial work between the stages of a compliance workflow: review an output, assess its relevance, decide what happens next, assign the task. The company’s attention had moved toward agents that could execute those connected steps, with people retaining oversight.
His argument stayed tied to the underlying data. An agent needs context about the actual organization, its controls and its operating criteria. A fluent answer about a framework is a different deliverable from an action grounded in a company’s current systems. His proposed path depends on connecting the action to evidence that someone can examine.
In a CISO Series discussion, he described agents acting on behalf of accountable people, with boundaries and traceability. He also resisted treating a passed audit as the whole objective. The larger concern was understanding and managing risk. That distinction puts responsibility back into the story: software can perform tasks while a person remains answerable for what happens.
Anecdotes’ 2026 GRC Data & AI Summit gave him an opening session titled “It’s the End of GRC as We Know It.” His more personal account of leadership was quieter. Reflecting on the company’s 2026 sales kickoff, he emphasized the daily decisions and difficult conversations he shares with Amior. The partnership, in his telling, takes its shape on ordinary working days.
That is where the million-hour ambition becomes interesting again. Every hour recovered belongs to somebody who can use it for a better question, a task requiring judgment, or work they wanted to do in the first place. Kuznitsov has spent his career expanding his own job beyond code. The business he is building asks how much of everybody else’s job the code can finally take care of.