When we hire people, even for narrowly defined roles, we invest time in onboarding. We explain how the organisation thinks, what matters, where judgment is expected, and where boundaries sit. Much of this has nothing to do with the task list itself.
This context is what allows someone to operate independently over time. Without it, even highly capable people struggle to make good decisions. With it, they begin to contribute in ways that go beyond their initial remit. Onboarding provides orientation before it produces results.
What we skip with AI
When it comes to AI, we often skip this step entirely. Systems are given access to tools and data and expected to deliver value immediately. When results fall short, the response is usually to adjust prompts or switch models.
What is missing is guidance.
AI is treated as something that should work out of the box, rather than something that needs to be situated inside a specific environment. Information is provided, but context is not. Instructions are given, but grounding is absent.
This knowledge gap becomes more visible as AI moves closer to daily work and decision support.
Governance as onboarding material
If AI is becoming a teammate rather than a tool, it needs onboarding in the same way humans do. Not only to provide rules and scripts, but to provide context. Context is what’s needed to enable better autonomy and decision-making.
This is where governance documents take on a different role. A family constitution, investment beliefs, or a statement of purpose is not an operational manual. They encode intent, time horizon, and the approach to trade-offs.
When these documents are brought into a knowledge environment, they become part of an AI system’s upbringing. They shape how questions are framed before answers are generated, and what considerations should remain present even when efficiency is tempting.
Most organizations, including family offices, already understand this instinctively when onboarding people. AI simply brings that instinct into sharper focus.