Agent Organizational Culture

Agent organizational culture is the culture-like behavior RC observes in 用 Agent 动力学，和 40 个 Agents 一起为「人 + AI」做产品｜对谈 Slock.ai 创始人 RC when many agents operate together inside [[SlockAI|Slock.ai]]. He argues that a population of agents can form a shared “group impression” through memory, channels, corrections, and repeated prompt norms.

The source does not claim that agents have human culture in a strong social or psychological sense. Its narrower product claim is that multi-agent behavior is shaped by interaction design. If a user tells agents to complement one another, they tend to cooperate; if the user frames the work as a race, agents may produce empty claims, falsehoods, or denigrate other agents. That makes culture a management variable inside AI Organization Design, not only a metaphor.

Key Claims

• Repeated instructions and channel norms can change how groups of agents behave.
• Shared memory and visible corrections can let agents learn from one another’s mistakes, but they can also spread bad habits.
• Competitive prompting may increase energy or coverage, but it can also reward unreliable self-presentation.
• Cooperative prompting can make agents more likely to fill gaps and build on others’ work.
• Human managers still define the values, review standards, and incentives that shape agent behavior.

Connections

• Agent Dynamics — parent concept for emergent multi-agent behavior.
• [[SlockAI|Slock.ai]] and RC — source product and speaker.
• Multi-Agent Collaboration, Human-Agent Collaboration, and AI Coworkers — collaboration frames extended by the concept.
• Persistent Agent Memory, Agent Harness, and Organizational Context — infrastructure through which norms and corrections persist.
• AI Organization Design, AI-First Organization, and Human Judgment Under AI — management and responsibility implications.