Google is undergoing a comprehensive "work style transformation" within the company. In recent weeks, Google's management has clearly informed some non-technical staff that they must use AI tools in their daily work, and the use of AI will be formally included in the annual performance evaluation criteria.

This move marks Google's shift of AI from an "optional add-on" to a "core productivity tool." Previously, Google required software engineers to use AI-assisted programming, and currently, about 50% of the code within the company is generated by AI and reviewed by humans.

AI usage requirements for different positions:

  • Software Engineers: Must use internal AI programming tools (such as Goose, based on Google's technical history training) to write code, solve technical problems, and maintain code repositories.

  • Non-Technical Positions: Required to use AI to write strategic documents, analyze sales calls, and extract customer insights.

  • Sales Team: Must use internal AI tools to record call notes, and some employees are required to reach specific AI usage counts per week, even using AI virtual avatar tools (such as Yoodli) for customer communication practice.

To ensure data security, Google employees mainly use customized internal AI tools. For example, Duckie, an internal version of Gemini, can answer questions based on the company's sensitive documents without worrying about leaks.

Google CEO Sundar Pichai stated that in light of the current widespread deployment of AI by competitors internally, Google must keep up. Currently, AI usage is linked to employee performance through the GRAD (Googler Reviews and Development) system. This means that the level of AI ability directly determines an employee's career development at Google.