First Week Onboarding Wizard

Replace the 38-SOP dump with 5 questions and focused task proposals. ~10 minutes per AI worker, not hours of system prompt engineering.

The Problem

Onboarding an AI agent today means writing system prompts, building Zapier zaps, configuring custom GPTs, and debugging what they actually do. There\u2019s no structured intake. The agent either overpromises (a 38-SOP dump on day one) or underdelivers (a single chat persona with no real scope). Managers spend hours configuring tools they then abandon.

How It Works

Answer 5 Questions (~10 min)

Priorities, Success Criteria, Guardrails, Collaborators, Cadence. AI pre-fills suggestions from position activities and KPIs.

AI Proposes 5–8 Focused Tasks

Not 38 generic SOPs. Each proposal is a one-paragraph pitch — what the task does, when it runs, what tools it needs, what success looks like.

Approve, Skip, or Edit Each

Manager reviews every proposal before any SOP is written. SOPs are written only for approved tasks.

Worker Starts with Context

Guardrails, collaborators, cadence, and success criteria are all baked into the worker’s system prompt and scope guardrails.

The 5 Questions

Priorities — What’s on fire

Pick the 2–3 position functions that matter most this week. Sets the scope of initial SOPs so the worker focuses where you need it.

Success Criteria — What good looks like

Per-function descriptions of what success looks like. Becomes the AI’s evaluation criteria and feeds the worker’s Activity History reviews.

Guardrails — What never to do

Free-text rules specific to this team and role. Baked into the AI’s system prompt and scope guardrails — not just a suggestion.

Collaborators — Who it works with

Pick the org chart positions this AI worker will interact with. Used by the AI to know who to escalate to, who to copy, who has authority.

Cadence — How to work with you

Briefing frequency, alert channel, response window. Sets the manager’s interaction rhythm.

Manager-in-the-Loop

Onboarding state persists. Re-onboard when the manager changes. Every AI worker has a human manager and a human escalation owner.

Who It's For

Managers deploying their first AI workerAI-First HR LeadersHR Directors