AI Coworkers

AI coworkers are role-based agents that participate in work as persistent collaborators rather than one-off answer boxes. In “AGI 来了？我用了一周，头皮发麻“｜对谈张昊然：Moxt 联合创始人, Moxt lets users create and train AI coworkers inside an AI-Native Workspace, with Momo as the default assistant.

The episode describes AI coworkers as having goals, memory, skills, workspace access, and communication surfaces. Zhang Haoran gives examples from his own setup: a manager-like agent that tracks project sync, a Golden Sales agent seeking 1000 paying users, agents for deep thinking and creative work, and a critic that checks whether the team and founder are focusing on the right issues.

This concept overlaps with Digital Employees but carries a different value posture in the Moxt source. Digital-employee language often frames AI as labor to deploy and evaluate, while Zhang Haoran explicitly objects to marketing AI as a low-cost human replacement. Moxt’s preferred framing is that AI coworkers should amplify human judgment, taste, feedback, and value choices.

Key Claims

• Multiple specialized agents may be easier for humans to manage than one all-purpose digital clone because each agent can have a clearer goal, context, skill set, and memory.
• AI coworkers require onboarding and feedback: users train them by telling them what was good, wrong, or misaligned.
• Cost can become a management target for the agent itself, not only a product-level billing concern.
• AI coworkers change the human role toward direction, review, feedback, approval, and meaning-making.
• The coworker metaphor increases the need for permission, attribution, privacy, and workspace-value boundaries.

Connections

• Moxt, Momo, and Zhang Haoran — product, default assistant, and founder source.
• Organizational Context — shared context that lets AI coworkers act with less manual briefing.
• Persistent Agent Memory, AI Skills, and Agentic Workflow — mechanisms that make coworkers trainable and repeatable.
• Multi-Agent Collaboration — possible pattern when multiple AI coworkers coordinate or critique each other.
• Human-Agent Collaboration, Proactive Agents, and Human Judgment Under AI — human-role and interaction implications.
• Digital Employees and AI Organization Design — adjacent enterprise and organization-design frames.