Israeli AI Startup Plurai Raises $13 Million for Automated Agent Testing
Translated & summarized from Calcalist by baba
Israeli AI startup Plurai has raised $13 million in seed funding to develop a platform that automatically tests and improves AI agents before deployment. Founded by experienced AI researchers, the company aims to address the challenge of transitioning AI agents from development to production. The funding will be used to expand R&D, grow its US presence, and enhance its automated testing and refinement capabilities for AI agents.
The story in 5 lines · by baba
- Israeli AI startup Plurai secured $13 million in seed funding to enhance AI agent reliability.
- The company's platform automatically tests and improves AI agents in simulated environments.
- Plurai was founded by Dr. Ilan Kedar and Dr. Elad Levy, with expertise in AI and autonomous systems.
- The funding will support R&D expansion and growth in the US market.
- Investors include Team8, Velocity, U&I, Mercer, and prominent angel investors.
Israeli artificial intelligence company Plurai has emerged from stealth mode, announcing a $13 million seed funding round. The company's platform is designed to automatically test and improve AI agents before they are deployed in real-world scenarios, and to continuously enhance their reliability and quality post-launch. The funding round was led by Team8, with participation from Velocity, U&I, and Mercer, alongside prominent angel investors. Among the investors are Oudi Mokady, founder of CyberArk, and Ria Miera, Senior Director of Google Models at Google Cloud and former head of model evaluation at OpenAI. Dr. Eyal Carmi, CTO of Microsoft 365 Copilot, serves on Plurai's advisory board.
Plurai was founded in 2025 by Dr. Ilan Kedar, who serves as CEO, and Dr. Elad Levy, who is CTO. Both have extensive experience in AI research and development, with doctorates in computer science and mathematics respectively. They previously worked together at Nexar, evaluating systems for autonomous vehicles like Waymo through simulations. The company recently announced a collaboration with NVIDIA and Nebius, showcasing technology that allows developers to adapt open-source AI models for specific tasks in about 20 minutes, achieving higher accuracy at a significantly lower cost compared to leading models from OpenAI and Anthropic.
Currently employing around 12 people in Israel and the US, Plurai plans to double its workforce by the end of 2026 and expand its US operations. The new funding will support the expansion of R&D teams, deepen engagement with enterprise clients, and accelerate US growth. A key challenge for organizations developing AI agents is transitioning them from demo to production while maintaining the trust needed for large-scale deployment. Plurai's solution, inspired by autonomous vehicle testing, creates a digital twin of the AI agent and its operating environment. It runs the agent through thousands of simulated scenarios, including rare edge cases, to identify functional failures, incorrect responses, unsafe actions, and unexpected behaviors before production. The platform then automatically refines the agent and re-tests it to ensure performance improvements without regressions. It also facilitates the replacement of expensive large language models with smaller, task-specific models to enhance accuracy and reduce operational costs.