French AI startup H Company is making waves in the artificial intelligence world with its new release of three agentic models. Launched just last year, H Company raised hundreds of millions of euros before unveiling its first products. Their agentic AI agents don’t just answer questions — they plan, execute tasks, and solve problems, marking a shift beyond traditional chatbots. The company insists humans stay central in the process, ensuring AI assists rather than replaces human decisions. This new development is another important AI milestone, especially within Europe’s strict data privacy environment.
What’s Happening & Why This Matters
On Tuesday, H Company launched three agentic AI models named Runner H, Tester H, and Surfer H. These models are designed to perform complex, real-world tasks with a 92.2% success rate. The company claims this efficiency comes with cost savings up to 5.5 times compared to rivals such as OpenAI, Anthropic, and Google.
Runner H automates workflows and streamlines tasks aimed at consumer users. It builds on concepts from other AI agents, such as China’s Manus, offering more advanced automation capabilities.
Surfer H excels at browsing the web and navigating online environments. It scored 92.2% accuracy on the WebVoyager benchmark, costing just $0.13 (€0.11) per task. By comparison, OpenAI’s Operator agent charges around $200 (€175) per month for its Pro plan. Surfer H acts on behalf of users by planning, recognizing web interfaces, clicking, scrolling, and gathering information efficiently.
Tester H targets enterprise use, specializing in software testing and quality assurance. It also supports smart email replies by drafting responses that users can review and send themselves, keeping humans in control.
All three models rely on synthetic data, which mimics real-world data while fully complying with the European Union’s GDPR data protection rules. This approach enables H Company to respect user privacy while advancing AI capabilities.
Charles Kantor, the company’s CEO and co-founder, shared his personal experience with the models, using them for tasks such as managing emails, preparing billing information, and crafting marketing content. He describes the AI’s output as drafts that he then reviews, underscoring the blend of AI productivity and human oversight.
H Company’s clear agentic vision is to build “trusted, action-oriented AI that delivers task execution beyond traditional chatbots.” Kantor believes agentic AI is the next wave of innovation, essential for robotics and simulated environments. He predicts the technology will become central to computer games, robotics, and virtual worlds, referring to agentic AI as the “heart” of many future applications.
Founded by Stanford and DeepMind alums, H Company has attracted investment from industry heavyweights, including LVMH’s Bernard Arnault, Iliad’s Xavier Niel, Amazon, Samsung, and former Google CEO Eric Schmidt. Although three co-founders left due to operational differences, Kantor describes the split as “in the past,” emphasizing a unified vision moving ahead.
With around 70 employees in Paris and the UK, H Company invests heavily in research and engineering. Kantor calls the organization’s approach “a concerto of age,” highlighting the need for cohesion across product, research, and engineering to build category-defining AI solutions.
TF Summary: What’s Next
H Company’s launch of Runner H, Tester H, and Surfer H places it at the forefront of the agentic AI race. Their models combine task execution with cost efficiency and strict compliance with GDPR rules, promising to reshape how AI assists in daily tasks and enterprise operations. This approach sets a strong foundation for agentic AI’s role in robotics, gaming, and web navigation.
The company’s leadership, backed by top-tier investors, points toward rapid growth and innovation in the AI agent space. With human oversight at its core, H Company balances advanced AI capabilities with ethical considerations. As agentic AI becomes more widely accepted by various industries, agent technology could redefine the relationship between humans and machines.
— Text-to-Speech (TTS) provided by gspeech.