What most AI onboarding platforms get wrong
AI onboarding isn't magic. It works because it starts from the actual role — not a template someone copied in 2019.
The checklist was never the problem. The problem was that the checklist was generic, handed to a manager who half-remembered the role requirements, built from a template that predated the current team structure, and completed without anyone checking whether the new hire actually understood any of it.
AI onboarding tools get pitched as a solution to the checklist. That's not quite right. The more accurate description: they're a solution to the person who had to build the checklist at 8pm the night before someone's first day — and who, pressed for time, just reused last year's version.
What AI actually changes
The real bottleneck in onboarding has always been manager bandwidth combined with role complexity. A good manager, with enough time, can design a thoughtful 90-day plan that's specific to the hire, the role, and the current state of the team. It would cover what success looks like at 30, 60, and 90 days. It would account for the fact that this hire is a senior engineer, not the junior one from last year. It would set up the right early conversations and the right early milestones.
Most managers don't have that time. So they reuse what already exists, skip the role-specific thinking, and hope the new hire figures out the gaps through osmosis. This isn't negligence — it's rational behaviour under constraint. Onboarding planning is important but non-urgent work, and non-urgent work gets deprioritised by people who are managing projects, running interviews, and handling their own deliverables.
AI onboarding doesn't replace manager judgment. It replaces manager effort on the structural work — the scaffolding that should have existed already but usually doesn't. Upload a job description and a CV, and you get a milestone plan that reflects the actual role. A senior hire gets different milestones than a junior one. A sales hire gets different 30-day goals than an operations hire. That differentiation used to require hours of focused work. Now it doesn't.
What managers then do with that plan — the conversations they have, the context they add, the relationships they help build — still comes from them. The AI gives them a starting point that's actually relevant, instead of a blank document or a recycled template from a role that no longer exists.
Why most platforms still miss
The category has matured enough that most HR platforms now offer some version of "AI onboarding." Most of them are applying AI to the wrong layer.
Automating email reminders isn't AI onboarding. Surfacing a library of generic learning resources isn't AI onboarding. Sending a new hire a welcome video with an AI-generated script is arguably worse than nothing — it signals that you've automated the communication without doing the underlying thinking.
The platforms that are getting this right are using AI at the planning layer: at the point where someone decides what this specific hire should accomplish in their first 30, 60, and 90 days. That's where the most time is lost and where the most value is created. A good plan makes every downstream step easier — the manager check-ins land better, the 30-day survey produces more meaningful data, the transition from "new hire" to "contributor" happens faster and with less friction.
A bad plan — or no plan — means the new hire spends their first month triangulating expectations from disparate sources and trying not to ask too many questions in case it makes them look unprepared. That's an exhausting way to start a job. It's also one of the most reliable predictors of early attrition.
The thing that won't change
There's a version of this conversation that ends with "AI will eventually handle all of it." It won't.
Onboarding is ultimately about a new person finding their footing in a specific team, with specific people, in a specific moment in that company's history. The structural work — the plan, the task list, the milestone tracking — can be generated and managed by software. The human work can't. The manager who takes 20 minutes on day three to explain not just what the new hire should be doing but why it matters right now. The teammate who flags the unwritten norms before the new hire violates them. The recognition, two weeks in, that tells someone their contribution was seen.
AI makes the structural work fast enough that managers have time for the human work. That's the actual value proposition. It's not that the technology is smart. It's that it removes the friction that was stopping people from doing their jobs well.
The On&On Advantage
On&On's AI onboarding starts at the planning layer. When you add a new hire, you upload their job description and CV — and On&On generates a personalised, milestone-based journey with specific goals and actions per phase. The manager gets one view, the employee gets another, and both know exactly what's supposed to happen and when. There's no template to customise, no library to browse. The plan is specific to this hire, in this role.
Most managers have the plan ready before the new hire's first day. That's the part that used to take a weekend.
On&On is $9 per employee per month with no lock-in. Most teams are set up the same day they sign up.
Ready to replace the generic checklist? Request a demo to see how On&On builds personalised onboarding in minutes.