Amazon has unveiled Nova Act, a general-purpose AI agent designed to control web browsers and perform independent actions. Alongside this advanced AI technology, Amazon is also releasing the Nova Act SDK, a toolkit for developers to build agent prototypes powered by Nova Act.
Developed at Amazon’s recently established San Francisco-based AGI lab, Nova Act is set to play a crucial role in the upcoming Alexa+ upgrade, an AI-enhanced version of Amazon’s voice assistant. However, the version available today is labeled as a research preview, indicating that it’s still in development.
Aiming to Compete in the AI Agent Space
Developers can now access the Nova Act SDK through nova.amazon.com, which also showcases Amazon’s Nova foundation models. The company’s move into AI agent technology is seen as a response to competitors like OpenAI’s Operator and Anthropic’s Computer Use, both of which aim to enable AI-driven web navigation.
While Amazon isn’t the first to develop such AI agents, its Alexa+ integration could allow Nova Act to reach a significantly larger audience than its rivals. Developers using the Nova Act SDK can automate everyday tasks such as ordering food from Sweetgreen or making dinner reservations. The toolkit enables AI-powered navigation, form filling, and calendar selection.
Amazon claims that Nova Act outperforms OpenAI and Anthropic in its internal tests. For instance, on the ScreenSpot Web Text benchmark—an evaluation of AI agents’ ability to process on-screen text—Nova Act scored 94%, compared to OpenAI’s CUA (88%) and Anthropic’s Claude 3.7 Sonnet (90%). However, Amazon has not yet benchmarked Nova Act against widely recognized agent tests such as WebVoyager.
A Product of Amazon’s AGI Lab
Nova Act is the first public product from Amazon’s AGI lab, co-led by former OpenAI researchers David Luan and Pieter Abbeel. Luan previously founded Adept, while Abbeel co-founded Covariant before Amazon recruited them last year to lead its AI agent initiatives.
Although automating simple tasks like ordering food may seem basic for an AGI-focused lab, Luan argues that AI agents represent a key step toward superintelligent AI. He defines AGI as an AI system capable of handling any task a human performs on a computer. The Nova Act SDK is designed to reliably automate short, simple tasks while allowing developers to specify when human intervention is needed in an agentic workflow.
Will Nova Act Deliver on Its Promise?
Amazon is betting big on AI agent technology, and Nova Act’s success could influence the future of Alexa+. The company is entering a competitive field, where early AI agents from OpenAI, Google, and Anthropic have struggled with reliability and autonomy across different domains. These systems tend to be slow, prone to errors, and require frequent human intervention.
With Nova Act, Amazon aims to overcome these challenges, but the true test lies ahead. Early trials of Nova Act will reveal whether Amazon has achieved a breakthrough—or if its AI agents will face the same limitations as its competitors.
Also Read : Mark Zuckerberg Reflects on Watching ‘The Social Network’ with Facebook Employees
