Stop Throwing New Hires Into the Deep End: The Case for an AI Onboarding Concierge
Every company talks about wanting to give new employees a great start. Most deliver a PDF checklist, a Slack invite, and a vague "let me know if you need anything."
Then they wonder why it takes 90 days for someone to feel productive.
RAG #2 in the Kendall Project's 100 Essential Enterprise AI Assistants tackles this head-on: the Onboarding Concierge.
The Problem No One Talks About
Starting a new job is overwhelming. Not because people aren't smart enough but because the information they need is scattered across a dozen systems, locked in colleagues' heads, and buried in onboarding docs that were last updated two years ago.
Ask yourself: when a new hire joins your team, can they independently answer these questions in their first week?
- What systems do I need access to, and how do I request them?
- What training is required before I can start work?
- What does my first 30/60/90 days actually look like for my role?
- Who do I need to meet, and how do I set that up?
If the answer is "they'd have to ask someone," you have an onboarding problem.
What the Onboarding Concierge Does
The Onboarding Concierge is a context-engineered AI assistant built on RAG (Retrieval-Augmented Generation) that guides new employees through their first 90 days with personalized guidance, not generic advice.
It adapts based on three key variables:
- Role - What does this person actually need to do their job?
- Department - What are the team-specific norms, tools, and workflows?
- Location - Are there region-specific compliance requirements, benefits rules, or office procedures?
The result is an assistant that can tell a new software engineer in Austin exactly which dev environment setup steps to follow, which compliance trainings are required before their first sprint, and who to contact for VPN access, without routing every question through a busy HR team or manager.
The Context Blocks That Make It Work
What separates the Onboarding Concierge from a fancier FAQ page is the structured context it's built on. The Kendall Framework identifies five critical context blocks for this RAG:
- Roles - What's expected of this person and what do they need to succeed?
- Processes - The actual step-by-step workflows for system access, training completion, etc.
- Assets - The tools, templates, and resources new hires need to find and use
- Locations - Office-specific, regional, or country-specific variations
- Glossary - The company-specific language and terminology that every insider knows but no one explains
Get these five blocks structured and maintained, and you have an AI that can genuinely answer a new hire's question at 9pm before their first day without anyone working late to help them.
The Business Case Is Simple
The cost of a slow or broken onboarding process compounds fast:
- Extended time-to-productivity is lost revenue and delayed impact
- Missed compliance training creates regulatory exposure
- A bad first-week experience increases early attrition and early attrition is expensive
The Onboarding Concierge doesn't replace a good manager or a welcoming team culture. It handles the operational complexity of getting someone set up, so managers can focus on the human side of bringing someone new onto the team.
Where to Start
If you're thinking about building an Onboarding Concierge for your organization, the first question isn't "what AI tool should we use?" It's: do we have our context structured?
Start by auditing the five context blocks above. For most organizations, Roles and Processes are the places where the biggest gaps live - either the documentation doesn't exist, or it exists but no one has updated it since the last reorg.
Fix the context first. The AI can follow.