Amazon Nova Act: A New Milestone in the Journey Toward Autonomous AI

Amazon Nova Act AI

The evolution of artificial intelligence is no longer just about generating content — it's about thinking, acting, and executing tasks independently. In a bold leap toward Artificial General Intelligence (AGI), Amazon has introduced Nova Act, a next-generation AI model developed by its AGI Lab in San Francisco.

Nova Act is more than a language model. It represents a new category of AI: autonomous agents capable of completing real-world tasks in web browsers without human intervention. This marks a significant paradigm shift in how we interact with and integrate AI into everyday digital environments.

What Is Nova Act and What Can It Do?

Nova Act is designed to understand natural language instructions and convert them into actions performed directly within a web browser. In practice, this means it can:

• Fill out online forms

• Make reservations or purchases

• Navigate websites and retrieve information

• Compare prices or complete multi-step processes

Unlike traditional chatbots or models that generate text, Nova Act can interpret web elements (like buttons, menus, and forms) and interact with them in real-time, making it a fully functional autonomous agent.

Benchmark Performance Amazon Nova Act AI: Leading the Pack

In internal benchmarking tests, Nova Act outperformed leading competitors such as OpenAI and Anthropic on key tasks, especially in:

• ScreenSpot Web Text

• GroundUI Web

These tests evaluate how well an agent can recognize interface elements, follow user goals, and execute tasks with minimal steps. Nova’s accuracy in navigating and interacting with web pages sets a new standard for real-world agent performance.

Amazon Nova Act AI SDK: Empowering Developers

To support innovation, Amazon has launched nova.amazon.com, offering access to the Nova Act SDK. With it, developers can:

• Build custom task agents

• Integrate Nova into web applications

• Experiment with autonomous workflows

This opens up opportunities for businesses to automate web-based tasks that previously required manual input — from operations and customer service to market research and testing.

Amazon’s AGI Vision: Shifting the Core of Computing

David Luan, Director of Amazon’s AGI Lab, emphasizes the transformative nature of Nova Act:

“In the future, compute calls will go directly to large AI agents.”

This vision reimagines the fundamental structure of computing — from passive instruction-based systems to intelligent agents capable of reasoning and autonomous execution. Nova Act represents an early prototype of this shift.

From Alexa to Nova: Expanding Everyday Applications

Nova’s integration into Amazon’s ecosystem could redefine the assistant experience. Imagine asking Alexa:

• “Book me a hotel for this weekend.”

• “Order the cheapest version of this product online.”

• “Fill out and submit this government form.”

Instead of reading search results or offering links, Alexa could delegate the task to Nova — who would actually go online and complete it for you. This is where virtual assistance transforms into digital agency.

 Conclusion: The Rise of AI Agents Has Begun

Nova Act is a powerful signal of where AI is headed. Beyond content generation, it steps boldly into the realm of real-world interaction. Its ability to understand, plan, and act — autonomously — positions it as a blueprint for the future of digital work.

As AGI ambitions grow, Amazon's Nova Act offers a glimpse of a not-so-distant future where intelligent agents navigate the web, solve problems, and execute tasks as true digital collaborators.

References:

• Amazon Science – Nova Act Introduction

• Wired – Amazon’s AI Agents Are Already Browsing the Web

• The Verge – Amazon Launches Nova SDK

• About Amazon – Nova Website & Tools