We hired an agent. It onboarded like everyone else.

An AI agent was integrated into an engineering team, where it performed tasks such as raising pull requests and filing tickets, much like a human colleague. Initially, it made typical onboarding mistakes, such as using incorrect labels and assigning tickets to non-existent teams, highlighting the need for explicit documentation of tribal knowledge and local rules. The AI agent's ability to self-correct and update its configuration based on feedback was compared to a graduate hire learning from mistakes, emphasizing the importance of publicly addressing errors to reinforce standards for both the agent and human observers. The narrative underscores the necessity of implementing safeguards, such as audit trails and permission restrictions, to ensure that AI agents operate safely within defined limits, while also noting the unsettling yet ordinary experience of collaborating with a non-human team member.