Zuckerberg Sets "Hard Targets" for Meta Employees: 80% of Senior and Mid-Level Engineers Must Use AI in Their Work

According to Business Insider, Mark Zuckerberg is working to transform Meta into a fully “AI-native” company. An internal document shows that Meta is pushing employees to use AI more deeply in their work by setting clear targets.

Meta has established specific requirements for certain roles, such as mandating heavy reliance on AI tools in tasks like programming. In addition, Meta employees have compiled AI usage targets for different business units, covering both goals set at the end of last year and planning toward 2026.

Among them, the Creation organization, which is responsible for core creative experiences, has proposed that by the first half of 2026, 65% of engineers are expected to use AI to write more than 75% of committed code (i.e., code that is saved and tracked in projects).

The Scalable Machine Learning organization, responsible for AI models and infrastructure, has set a target for February 2026 for 50% to 80% of code to be AI-assisted. However, a senior engineering manager added, “We will not track this by metrics.”

At the company level, Meta has also set unified goals for its core product team for the fourth quarter of 2025. This team covers products including Messenger, WhatsApp, and Facebook. One goal is to ensure that 80% of mid-level and senior engineers use AI tools like DevMate, Metamate, and Google’s Gemini, while emphasizing that the main focus is on the adoption rate of these tools, rather than the proportion of AI-generated code.

In addition, 55% of code changes made by software engineers in the core product team must involve participation from AI agents.

It is still unclear whether these targets will directly affect performance evaluations.

A Meta spokesperson stated, “This is a clear priority direction, and we are promoting the use of AI to improve the efficiency of employees’ daily work.” The spokesperson also emphasized that the performance system places more value on the real-world outcomes AI brings, not just the frequency of use.

Zuckerberg is accelerating this transformation. Business Insider previously reported that Meta has already started incorporating AI usage into the performance system, and employees are also using internal AI tools to write evaluation feedback for colleagues.

Recently, Meta has also introduced new title structures in some Reality Labs teams, including “AI Builder,” “AI Pod Lead,” and “AI Org Lead.”

These adjustments are taking place as Meta shifts toward a small-team model, compressing management layers and moving to a flatter organizational structure.

An internal memo writes: “Our ultimate objective is to achieve leaps in engineering efficiency and product quality. To this end, we are fundamentally restructuring the way we operate, our organizational structure, and our collaboration models.”

Meta CTO Andrew Bosworth announced on Tuesday that he will be responsible for the “AI for Work” project and promote the widespread adoption of internal AI tools. Meanwhile, this week Meta laid off hundreds of employees in Reality Labs and other departments.