Resemble AI has launched Resemble Agent Detection, powered by its RAD-S1 behavioral model, to identify AI agents that appear human on websites and in products. The system classifies each session as a person, bot, or AI agent by analyzing cursor movement, typing rhythm, on-page actions, and environment signals rather than relying on claimed identity or traditional bot fingerprints. In internal testing across multiple website stacks, RAD-S1 detected at least 98.8% of agents designed to conceal their automation, including those using stealth browsers. Six widely used bot detection tools identified only 3.1% to 14.8% of the same traffic. The product does not block sessions itself. Customers decide how to respond, such as allowing agents, redirecting them, or restricting sensitive actions. The launch follows moves by major banks to limit AI agents due to liability concerns and a series of agent-related security incidents.
Resemble AI has launched Resemble Agent Detection, powered by its RAD-S1 behavioral model, to identify AI agents that appear human on websites and in products. The system classifies each session as a person, bot, or AI agent by analyzing cursor movement, typing rhythm, on-page actions, and environment signals rather than relying on claimed identity or traditional bot fingerprints. In internal testing across multiple website stacks, RAD-S1 detected at least 98.8% of agents designed to conceal their automation, including those using stealth browsers. Six widely used bot detection tools identified only 3.1% to 14.8% of the same traffic. The product does not block sessions itself. Customers decide how to respond, such as allowing agents, redirecting them, or restricting sensitive actions. The launch follows moves by major banks to limit AI agents due to liability concerns and a series of agent-related security incidents.