The parts of onboarding AI genuinely speeds up, the parts it should never touch, and a week-one workflow I hand to managers.
In my consulting work, onboarding is the process teams most consistently underrate and AI most reliably improves. Not the human parts, the scaffolding: the 30-60-90 plan, the welcome guide, the tool one-pager, the check-in questions. Draft those with AI, spend the reclaimed hours on the moments a new hire actually remembers, and you close the gap most companies never do.
After years of helping teams put AI to work, I’ve noticed something. Companies pour their new AI tools into marketing and sales, the visible stuff, and walk straight past the place with the fastest payback: the first few weeks of a new hire’s job. Onboarding is unglamorous. It’s also the exact window where a good hire quietly decides whether they made a mistake.
Let me give you the number I open every HR workshop with. Gallup found only 12% of employees strongly agree their organization does a great job onboarding new people.[1] Twelve. That’s not a talent gap or a training gap. It’s a process gap, and process gaps are what AI was born to close.
Look at where the time actually goes. Someone has to write the welcome doc, build the first-week schedule, draft the intro to the team, list the tools and logins, and remember to book the 30-day check-in. None of that is hard. All of it slips when you’re covering the role and hiring for it at the same time. That slippage is the gap. AI closes it by turning “I’ll write that up later” into a solid draft you edit in ten minutes.
And this isn’t just about being welcoming. Gallup also found that people who had an exceptional onboarding experience are 2.6 times as likely to be extremely satisfied with their workplace.[1] The upside is large and the effort, with a bit of help, is small. That’s about as good as a business case gets.
Before we touch a prompt, let me draw the line clearly, because the fastest way to make onboarding worse is to hand the human parts to a chatbot.
Lean on AI for: the 30-60-90 plan, turning a messy pile of notes into a clean welcome guide, role-specific first-week checklists, manager check-in questions, and rewriting your dense internal docs into something a nervous new starter can read on day one without a translator.
Keep AI away from: the relationship. The manager who remembers your kid’s name. The teammate who takes you for coffee and tells you who really makes the decisions. The honest “how are you finding it” chat in week three. Automate those and people can tell, and what they feel is that they’ve joined a company that couldn’t be bothered.
One limit people underestimate: AI doesn’t know your company. It’ll invent a benefits policy or reference a tool you don’t use, and it’ll do it with complete confidence. Anything it drafts about your actual processes needs a human to check against reality before a new starter ever sees it. Treat the output as a strong first draft, never a source of truth. If you want to build the habit of catching that, our guide on testing AI prompts before you trust them is the discipline I teach.
Start here, because the 30-60-90 day plan is the highest-value document AI can build for you and the one managers most often skip. I’ve lost count of the number of “plans” that turned out to be a vague idea in one busy person’s head.
The whole trick is context. Give the model enough that the plan sounds like your company rather than a template it could hand anyone. Feed it the job description (if yours is weak, our post on writing job descriptions with ChatGPT fixes that first), the team’s real priorities this quarter, and the three things you most want this person doing well by month three.
Then insist on phases. Days 1 to 30 for learning and low-stakes wins. Days 31 to 60 for owning a small slice of real work. Days 61 to 90 for running with it. I push managers hard on this framing because the common failure is loading week one until the new person feels behind before they’ve found the bathroom.
A well-briefed plan comes back with milestones, the meetings that matter, learning resources, and a first project sized for someone still finding their feet. You’ll rewrite maybe a third of it, and that’s exactly the win. Editing a draft is a fifteen-minute job. Facing a blank page is the thing that never happens.
These are the ones I actually reach for, cleaned up so you can paste and adapt. Swap the bracketed bits for your details.
“You’re an experienced people manager. Draft a 30-60-90 day onboarding plan for a new [job title] joining a [team size] [industry] team. Their main goal by day 90 is [outcome]. Break it into learning, contributing, and owning phases. Include weekly milestones, who they should meet, and one starter project for weeks 3 to 4. Keep it realistic, not aspirational.”
“Turn the notes below into a friendly, skimmable welcome guide for a new starter. Short sections, plain language. Assume they know none of our internal jargon and explain acronyms once. Flag anything that looks out of date so I can double-check it. Notes: [paste].”
“Give me six open questions for a new hire’s first 30-day check-in. I want to surface whether they feel set up to succeed, where they’re stuck, and anything we’ve done badly. Warm and direct, not a survey. No yes/no questions.”
“Rewrite this list of tools and logins into a clear one-pager for someone on day one. Group by what they’ll use daily, weekly, and rarely. One line each on what it’s for. List: [paste].”
If your team keeps rewriting the same prompts, put them somewhere shared. I walk through exactly how in our guide to building an AI prompt library, and there’s a ready-made set in our ChatGPT prompts for HR collection.
Generic onboarding is the kind that makes a person feel like a headcount. What I love about doing this with AI is that personalising costs almost nothing once your base is built.
Keep one master onboarding doc and one master 30-60-90 prompt. For each new hire, change three variables: their role, their manager’s priorities this quarter, and their experience level. A senior leader joining to run a function needs a very different first month from a graduate in their first job, and the model reweights the whole plan in a single pass. Same scaffolding, right emphasis.
A small move that lands out of proportion to the effort: tell the model the tone of your workplace. Someone walking into a fast, informal startup and someone walking into a careful, regulated finance team both deserve a welcome that sounds like the place they actually joined. Name which one it is and let the AI match it.
The question I get most is which tool to buy. Nine times out of ten my answer is: none, not yet. A capable general assistant like ChatGPT, Claude, or Gemini handles every drafting task in this article with no specialist HR platform at all. Start there, prove the workflow saves you hours, and only then consider spending.
A purpose-built onboarding tool starts to earn its price at scale. If you’re hiring ten people a month across several teams, a system that assigns tasks, tracks completion, and nudges managers automatically is worth it, because at that volume the failure isn’t writing the plan, it’s remembering to run it for everyone. Below that, a shared doc and a strong set of prompts will outperform most software and cost nothing.
What I’d steer you away from is anything promising to “automate the new-hire experience” end to end. That phrase should make you flinch. The experience is the human part, and a tool that automates it is proudly optimising for the wrong thing. Buy tools that remove admin, not tools that remove people. And whatever you pick, keep your prompts and your best plans somewhere that outlives the subscription, so changing vendors later doesn’t mean starting from scratch.
Remote onboarding is where the small gaps turn into big ones. A new starter at home can’t lean over and ask the person beside them what a term means or whether it’s normal that a system is down. Everything that was tacit in an office now has to be written down, which is exactly the documentation nobody ever gets round to.
This is a strong use of AI. Ask it to turn your scattered process notes into a proper remote-starter guide: how you communicate and on which channel, your meeting norms, who to message when you’re stuck, and the unwritten rules an in-office hire would absorb by osmosis. Then have it draft a lightly structured first week that front-loads real human contact, short video calls with the people they’ll actually work with, rather than a wall of documents to read alone in silence.
One thing I always add for remote hires, and no tool can do it for you: an explicit “it’s completely normal to feel lost in week one” line, and a named person whose job that week is to field the questions that feel too small to raise. AI drafts the guide beautifully. It can’t be that person. Assign one anyway, because for someone onboarding from their spare room, knowing who to ask is worth more than any document you’ll ever produce.
I said it up top and it earns its own section, because this is precisely where I watch teams get it wrong. AI should never become the thing a new hire talks to instead of a person.
Three moments stay human, no exceptions. The welcome: a short, genuine note from the manager on day one, written by the manager, beats any beautifully formatted guide. The check-in: use AI to prep the questions, then have the real conversation and actually listen to the answers. And feedback in the first 90 days, when someone is most likely to quietly conclude it isn’t working. That’s a judgment call, not a template.
There’s a compliance layer too. Anything touching pay, benefits, contracts, or legal policy needs sign-off from a person who knows the current rules before it reaches a new starter, because AI will state an outdated policy with total confidence and no warning label. You’re the safeguard. If you’re rolling AI into people processes more broadly, the guardrails in our piece on using AI for performance reviews carry straight over.
Here’s the whole thing as a sequence, roughly what I’d run the week before someone starts. It takes about an hour instead of an afternoon.
Begin with the plan, since everything else hangs off it. Build the welcome guide, the tool one-pager, and the intro message from there. Book the check-ins now, while they’re front of mind, so they don’t quietly evaporate. Then on day one, do the single thing no tool will ever do for you: show up, say hello properly, and mean it.
The steps below are the version I hand managers who ask where to start. Run it once, save every output, and your second hire takes half the time. By your fifth, onboarding stops being the thing you dread and starts being a small competitive advantage.
No, and you wouldn’t want it to. AI is excellent at the scaffolding: the 30-60-90 plan, welcome docs, tool guides, and check-in questions. But the moments a new hire actually remembers, the personal welcome, the honest check-in, the first real feedback, need a human. Automate the busywork, protect the relationship.
Drafting the 30-60-90 day plan. It’s the highest-value document and the one managers most often skip because writing it from scratch is a chore. Give AI the role, the team’s priorities, and your day-90 goal, and you’ll get a solid draft to edit in minutes.
Only as a first draft. AI doesn’t know your company and will state outdated or invented policies with confidence. Anything about pay, benefits, contracts, or legal rules must be checked by a person who knows the current details before it reaches a new hire.
Keep one master plan and prompt, then change three variables per hire: their role, their manager’s priorities this quarter, and their experience level. Also tell the AI the tone of your workplace so the welcome sounds like the place they’re actually joining.
Any capable general assistant like ChatGPT, Claude, or Gemini handles all of this well; you don’t need a specialist HR tool to start. Pick the one your team already uses, save your best prompts somewhere shared, and build a small library as you go.
This article is part of Future Factors’ practical AI series for HR and people teams. It’s written by an AI educator who runs onboarding and enablement workshops, and focuses on using everyday AI tools to speed up onboarding without losing the human touch.
Weekly tips. Real examples. Practical help for busy professionals.
We care about your data, check out our privacy policy.