White Circle: Denis Shilov Raises €9.35 Million

White Circle Secures $11 Million Funding for AI Model Supervision

White Circle, a startup specializing in the supervision of artificial intelligence models, recently announced a funding round of $11 million, approximately €9.35 million. This financing was backed by several influential figures in the AI ecosystem, including Romain Huet, Dirk Kingma, Guillaume Lample, Thomas Wolf, François Chollet, Olivier Pomel, and Paige Bailey.

Founded by Denis Shilov, White Circle offers an innovative platform that allows companies to monitor the behavior of their AI models and autonomous agents in real time. The startup claims to already process over one billion API requests and counts among its clients international giants, including two major digital banks and the company Lovable.

This funding comes at a time when companies are ramping up the deployment of AI agents in critical environments such as customer relations, finance, cybersecurity, human resources, and software automation. The market is now shifting towards the operational supervision of AI models.

Denis Shilov gained recognition in 2024 by publishing a "universal jailbreak" capable of bypassing the security mechanisms of leading generative models. This prompt allowed users to obtain responses that were typically blocked by systems like those from OpenAI or Anthropic, quickly attracting the attention of several American AI labs. Following this, Shilov was invited to participate in Anthropic's bug bounty program before launching White Circle.

White Circle's platform positions itself as an observability and control layer between AI models and business applications. It analyzes the inputs and outputs of models in real time to detect anomalies such as hallucinations, prompt injections, behavioral drifts, and to block malicious actions and prevent leaks of sensitive data.

Companies can define their own control policies to determine what is allowed or prohibited. White Circle also offers automatic mechanisms for limiting, blocking, or banning actions, and operates with various AI model providers through a single API. The company claims to support over 150 languages.

In the use cases described by White Circle, the platform can prevent an AI agent from executing destructive commands, detect abnormal behaviors in financial workflows, or identify attempts to manipulate internal model rules.

The development of White Circle illustrates a rapid evolution in the generative AI market. Following the race for models and GPU infrastructures, a new segment is emerging around the supervision of AI systems in production. The widespread adoption of low-code tools and "vibe coding" platforms accelerates the deployment of AI applications in businesses, enabling non-specialized teams to connect models to databases, CRMs, ERPs, or financial tools within hours.

However, this democratization creates a new risk surface. Companies must now manage agents capable of interacting with critical systems without always having clear visibility into their actual behaviors. Risks include exposure of sensitive data, operational drifts, malicious manipulations, or the execution of unforeseen actions.

White Circle is precisely positioned on this layer of operational supervision, similar to companies like Datadog or Sentry that have established themselves in the cloud domain. The startup also seeks to bolster its technical credibility by publishing research on the risks associated with generative models. In 2025, it published "CircleGuardBench," a benchmark for assessing the robustness of AI moderation models under real-world conditions, and more recently "KillBench," a study based on over a million experiments conducted on fifteen models from OpenAI, Google, Anthropic, and xAI.

According to White Circle, these studies revealed behavioral biases related to nationality, religion, physical appearance, or certain cultural markers. The study also claims that some structured formats used in enterprise AI integrations significantly reduce the built-in refusal mechanisms of the models.

White Circle's funding comes amid a gradual tightening of regulatory requirements surrounding artificial intelligence. Companies must now demonstrate their ability to trace model decisions, control agent actions, document drifts, and limit legal risks associated with automation.

With the emergence of AI agents capable of directly interacting with operational systems, the behavioral supervision of models could become a structural component of enterprise AI architectures. White Circle is betting precisely on this evolution.