The Onboarding Decisions That Shape a Team Six Months Later
Most teams judge onboarding by how fast someone gets productive. Laptop on day one, access by day three, first ticket closed in week one. Those are useful signals, but they measure the wrong thing. The decisions that shape how a team performs six months later are made in the first two weeks, and most of them are made by default.
The first is what a new person learns "good" looks like. Nobody hands them a definition. They watch. If the first review they sit in waves through weak work to protect a deadline, that becomes the standard, whatever the quality guidelines say.
The second is who they go to. New joiners build their map of the team quickly: who answers questions, who decides, who is safe to disagree with. Left to chance, that map often routes around the people you most need them to work with.
The third, and the newest, is how they use AI. Most onboarding says nothing about it. So people form their own assumptions in the first fortnight: that it is encouraged everywhere, quietly frowned on, or fine as long as nobody asks. Six months later you have a team using the same tools in five different ways, with five different ideas of what needs checking.
Multiply that across an organisation and the cost becomes visible. In a previous organisation, delivery managers were onboarded within each account group. The company had global delivery standards, but how they reached a new hire depended on their leader: how much time that leader gave it, how complete and how consistent they were. Over time we ended up with many ways of working, not all of them aligned with the standards. Catching up took real investment: a community for each discipline to spread the standards, local discipline leads in every location, and compulsory training on the guidelines that carried legal consequences. I was one of the community leaders and co-wrote the global standards for my discipline, so I saw first-hand how much harder it is to correct habits than to set them. All of it was catch-up work, and far more expensive than getting onboarding consistent in the first place.
None of this needs a longer onboarding programme. It needs three deliberate choices, made the same way by every leader who onboards someone. Show new joiners real work and say out loud what made it good enough. Introduce them to the people who decide, not just the people who are friendly. And tell them plainly where AI is expected, where it is not, what must always be reviewed by a person, and who that person is.
Onboarding feels like an administrative task. It is one of the few moments a leader gets to set the team's habits before they set themselves.
I publish a new article on transformation, delivery and leadership every week here. For AI, subscribe to Decoded, my monthly newsletter where I decipher the news and what it means for leaders.
