A 30/60/90 onboarding plan drafted for every new hire

A role-specific first-90-days plan, drafted in minutes from the job description and your org context, so every new hire ramps fast on a real plan instead of figuring it out alone.

What you'll have when you're done

A 30/60/90 plan template that produces a genuine, role-specific ramp for each hire: weekly milestones, who they should meet, what "success at day 90" looks like, and the handful of documents they must read. The manager edits it in a few minutes instead of authoring it from scratch (which usually means it never gets written). Load the milestones into your HR system and the check-ins fire on their own.

You spend weeks hiring well and then onboard people badly

Here is the quiet waste in most companies. You run a careful hiring process, land a great person, and then onboarding is a laptop, a Slack invite, and "let us know if you have questions." The new hire ramps slowly, feels unmoored, and a chunk of your hiring investment leaks away in the first month. I have been the CEO who spent six weeks obsessing over a hire and then gave them a worse first day than a coffee shop gives a barista. The asymmetry is embarrassing when you say it out loud: enormous effort to get the person, near-zero effort to land them. Strong onboarding is the difference between a hire who is contributing at day 90 and one who is still guessing.

The reason it gets skipped is that a good 30/60/90 is real work to write, per role, every time. That is exactly the kind of from-the-blank-page drafting AI removes. Give it the job description and your context and it produces a structured plan the manager refines, so the plan actually exists. The stakes here are lower than the rest of the hiring set (this is post-hire), but the same rule holds: AI drafts the logistics, humans own the decisions.

What you need first

• The final job description for the role (ideally the outcome-based one).
• The hiring manager's goals for the person's first quarter.
• Your tools and process docs: what systems they need, key wikis, the canonical docs.
• A Claude Project with your org context loaded.
• Your HRIS (Rippling, BambooHR, Gusto) to fire the milestones automatically, optional but high-leverage.

Step-by-step

Step 1Give the AI the role and the context

A generic 30/60/90 is worthless; a role-specific one is gold. Feed the model what makes it specific:

Draft a 30/60/90 onboarding plan for [role]. Inputs:
- Job description: [paste]
- The manager's goals for this person's first quarter: [paste]
- Our tools/process they'll need: [list]
For each phase (days 1-30, 31-60, 61-90): weekly milestones, who they should
meet and why, what "success" looks like at the end of the phase, and the 3-5
documents they must read. Make it specific to this role, not generic.

Here is the shape of the first phase, illustrative, for a Customer Success Lead:

Days 1-30 · Learn the system and the customers

• Week 1: finish tooling setup (Zendesk, the CRM, read access to the data warehouse); 1:1 with each of the 6 direct reports; read the top-20-accounts brief.
• Week 2: shadow 5 customer calls across the health spectrum (2 thriving, 2 at-risk, 1 churned-and-won-back).
• Weeks 3-4: co-run 3 renewal conversations with a senior CSM in the room.
• Success at day 30: can name the top 20 accounts, their ARR, and their health without notes; has a written read on the two biggest at-risk accounts.
• Must-read: the CS playbook, last quarter's churn post-mortem, the product roadmap one-pager.

That is the level of specificity that makes a plan usable. "Get to know the team and the product" is a vibe; "can name the top 20 accounts and their health by day 30" is a target the hire and the manager can both check against. The manager's job in the next step is to argue with these specifics, not to invent them from a blank page.

Step 2The manager edits and owns it

The draft is a starting point, not the final plan. The hiring manager reads it, adjusts the milestones to reality, and takes ownership, because a plan the manager did not buy into is a plan that does not happen. This is the human-judgment layer: the AI got you to a complete draft in two minutes; the manager makes it real in ten.

Step 3Load the milestones into your HR system

Take the agreed milestones and check-ins and load them into your HRIS. Modern systems (Rippling especially) can fire onboarding sequences and reminders automatically, so the day-7, day-30, and day-90 check-ins happen on schedule instead of being remembered. This is what turns a document into a process.

Concretely, in a system like Rippling or BambooHR you build an onboarding workflow keyed to the hire's start date: day -3 provisions accounts and orders the laptop, day 1 fires the welcome sequence and assigns the must-read docs as trackable tasks, days 7 / 30 / 60 / 90 auto-create the check-in meeting and send the self-assessment form. The manager gets a reminder, not a blank calendar. The milestones from your 30/60/90 stop living in a doc nobody reopens and become dated tasks that nudge the right person on the right day. You build the workflow once per role and clone it for the next hire into that role, so the second hire's onboarding is nearly free to set up.

Step 4Keep a human on the check-ins

Automate the scheduling of check-ins, not the reading of them. When a day-30 survey or self-assessment comes back, a human (the manager) reads it and acts. Do not let an automated system draw conclusions about a person's performance from a form, and never let AI touch compensation or promotion criteria. The system handles logistics; people handle people.

Step 5Reuse and improve the template

After a few hires, you will see what consistently needs editing in the AI draft. Fold those fixes into your Project instructions so the next draft starts closer to right. The plan compounds: each hire makes the template better.

How you'll know it's working

New hires ramp faster and feel oriented instead of lost, ask your day-30 hires whether they knew what success looked like, and watch the answer shift to yes. The operational tell is that onboarding plans actually exist for every role now, instead of the strongest-willed managers writing them and everyone else winging it.

When it breaks

• The plan is generic. You did not give it enough role context. Feed it the real JD and the manager's actual goals, not a one-line description.
• The manager ignores it. They did not own it (Step 2). The draft is the AI's; the plan has to become the manager's.
• Check-ins still get missed. They are not loaded into the HRIS yet, or are sitting in someone's memory. Automate the scheduling.
• It tries to set comp or promotion bars. Cut that. AI drafts the ramp; people make the calls that affect someone's livelihood.
• Every hire gets the same plan regardless of level. You reused one template across a coordinator and a VP. Their ramps are not the same shape; keep the Project instruction role-aware and feed the actual JD each time.
• The day-90 success criteria are fuzzy. "Is comfortable with the tools" is not observable. Push the model for criteria a manager could check off ("can run the weekly forecast unaided," "has closed 3 deals end to end"), the same specificity that made the day-30 bar useful.

Make it yours. A remote hire needs the plan to over-index on deliberate introductions and written context, because the hallway osmosis that orients an in-office hire does not exist; build explicit "meet X, here's why" steps into week 1. And the trickiest case is the first hire into a brand-new function, where there is no manager who owns the ramp. There, you own it: the same workflow drafts the plan, but the day-30 and day-60 check-ins land on your calendar, because no one else is positioned to catch a first-in-function hire drifting.

Where this fits in your harness

This is the last stage of the hiring pipeline and the handoff into your team. It pays off the work upstream: a sharp job description and a rigorous resume screen and interview process land you a great hire, and a real 30/60/90 makes sure that hire succeeds. It also connects to how you run the team day to day: the same AI that onboards people helps you roll out AI to a non-technical team.