AI-onboarding: two things that are actually one

AI-onboarding means two things, and most people searching the term only know which one they mean.

The first meaning is onboarding a human with AI's help. A new hire who can ask their way through instead of reading a handbook or waiting for someone to have time to explain.

The second is onboarding an AI agent into the company. Giving an agent the same knowledge about how you work, what you stand for, and why you do things the way you do, so it can be used for real work instead of writing generic emails.

These aren't two projects that happen to share a name. They're the same project.

Why they're the same thing

A new hire and an AI agent are missing the exact same thing when they start: the context that was never written down. Not facts, those already sit in your systems. What's missing is the process, the judgment, and the history, the knowledge that only lives in the head of whoever built the company.

Build a structured source for that knowledge, and you've solved both problems at once. A new hire onboarded against that source gets up to speed in weeks rather than months. An AI agent trained on the same source answers with your logic instead of guessing from what's typical across the industry.

What it isn't

It isn't a course tool. Generating courses and quizzes from content you already have solves the wrong problem, because it assumes the knowledge is already written down somewhere. Usually it isn't. The hard part was never packaging knowledge, it's getting it out of the person who has it and never put it into words.

It also isn't an HR project first, even though the result shows up fastest in how quickly a new hire gets productive. The core is the same regardless of who's on the receiving end, a person or an agent: one shared source that replaces heads and scattered documents.

Where to start

Ask yourself whether a new hire, without asking anyone, could explain why you do something a particular way instead of the obvious alternative. If they couldn't, that knowledge only exists in one person's head, and it's exactly the knowledge that makes an AI agent generic too.

We build that structured source for companies like Sakura, and in Sweden the project may also qualify for the digitalisation grant, since it's digitalising business processes, not marketing or sales.