Plurai raises $13 million and opens its AI agent simulation platform
The AMW Read
The seed round and public launch add a company-level update to the agent testing market, without independent evidence of broader impact.
Plurai raises $13 million and opens its AI agent simulation platform
Plurai announced a $13 million seed round and general availability of its AI agent simulation platform on November 11, 2026. Team8 led the round, with participation from U&I, Velocity, Mercer and more than 20 angel investors and industry executives. Plurai also introduced a free offering for a product previously available to selected partners and customers.
The launch adds a player to the market for testing and monitoring AI agents. Plurai says its platform builds a software replica of an agent and its operating environment using the customer's knowledge base, tools and rules. It then tests the agent against simulated users, including hostile ones, to find problems such as policy violations, incorrect actions, failed tool calls and sensitive data leaks. Simulated services, documents and system state can be reset between runs, while production traces can inform later scenarios. These capabilities address a practical barrier to agent deployment: teams need to understand failures across actions and environments, not just assess a model's answers.
For builders, the proposed workflow connects pre-deployment tests with failures observed in production. Plurai says it traces failures to evidence and code locations, generates fixes and validates them against regression tests. Simulation data can also provide examples for additional training of small models used to run agents or evaluate their behavior. Investors should assess whether those tests reproduce meaningful production conditions and whether generated fixes reduce failures without introducing new ones; the source provides no independent performance results.
