Google restricts access to popular AI tool Agent Smith
▼ Summary
– Google employees are using a new internal AI tool called “Agent Smith” that automates tasks like coding.
– The tool became so popular that access had to be restricted due to high demand from employees.
– Agent Smith works asynchronously in the background and can be instructed via employees’ phones and internal chat.
– Google leadership, including cofounder Sergey Brin, emphasizes that AI agents are a major focus and expected for employee use.
– Employee adoption of AI tools at Google is being factored into performance reviews in some cases.
A new internal AI tool has quickly captured the attention of Google’s workforce. Known internally as Agent Smith, this platform is designed to automate complex tasks like coding and has surged in popularity to the point where the company had to restrict access to manage overwhelming demand. The tool represents a significant step in Google’s broader push toward agentic AI, systems that can plan and execute workflows with minimal human oversight.
Building upon the existing Antigravity coding platform, Agent Smith can interact with a suite of internal Google tools. A key feature is its asynchronous operation, allowing it to work in the background without requiring an active laptop connection. Employees can check its status and issue new instructions directly from their phones. For many software engineers, it has already become an indispensable resource, arriving as Google aggressively promotes AI adoption to enhance company-wide efficiency.
The launch aligns with a clear directive from leadership. Google cofounder Sergey Brin recently told employees that AI agents will play a major role at the company this year, signaling a strategic priority. This sentiment is echoed across the industry, with Meta CEO Mark Zuckerberg reportedly building a similar agent to assist in running his company. When asked about Agent Smith, a Google spokesperson stated the company is continually experimenting with agents to solve real-world problems but declined to share specific details.
What sets Agent Smith apart from earlier coding assistants is its level of autonomous execution and its integration into daily workflows. Because it can access employee profiles, it can automatically retrieve documents that would otherwise require manual searching. Furthermore, its availability through Google’s internal chat platform makes it uniquely accessible, streamlining the process for employees to delegate tasks.
Brin emphasized the growing importance of these technologies during a town hall in early March, hinting at development on a tool comparable to something called OpenClaw. It remains unclear if he was referring directly to Agent Smith or a separate project. The meeting included a lighthearted moment where Google’s business chief, Philipp Schindler, joked about recognizing when Brin’s own AI agent was responding to messages for him.
The push for AI integration is being driven from the top. CEO Sundar Pichai has made it clear that adopting AI internally is critical, arguing that competitors are doing the same. This expectation is now being formalized; some engineers were told last year that using AI for coding was an expectation, and more recently, employees in non-technical roles have received similar mandates. In certain cases, AI adoption is becoming a factor in performance reviews.
This top-down pressure is complemented by grassroots initiatives. For instance, Google’s infrastructure organization is running Project EAT, an internal effort focused on improving and standardizing how AI tools are deployed across the company. Together, these moves underscore a fundamental shift, positioning AI agents not as optional aids but as central components of Google’s operational future.
(Source: Business Insider)
