Amazon has introduced a new generative AI model called Nova Act, capable of performing actions in a web browser—an important step in the development of AI agent systems. Nova Act, presented as a research version, enables developers to create AI agents that can automate tasks such as filling out forms, navigating web pages, and managing workflows.
With Nova Act, Amazon aims to ensure reliability in executing complex, multi-step tasks. The model breaks workflows into atomic commands, allowing for precise instructions and integration with tools for direct browser manipulation. Early tests show that Nova Act outperforms competitors such as OpenAI and Anthropic in certain web content interaction tasks.
The Nova Act model is also part of an upcoming Alexa+ update, making it the first public release from Amazon AGI’s lab in San Francisco. This lab aims to advance artificial general intelligence by training agents in diverse environments. Nova Act is positioned as a foundational step toward creating more autonomous and reliable AI systems, with potential applications ranging from IT automation to consumer services.
Amazon’s entry into the AI agent arena coincides with efforts from competitors like OpenAI and Anthropic, but emphasizes reliability and developer control. While Nova Act is still in its early stages, it could play a key role in shaping the future of task automation with AI.
