AI employee onboarding assistant
A professional-services business where this lands in HR.
02 — What is actually going wrong
Every new hire spends their first fortnight asking colleagues questions that are already answered somewhere in the intranet. The answers are inconsistent because they come from whoever was free, and the senior people being interrupted are the expensive ones.
How it works
An assistant grounded in your actual policies, system guides and process documentation answers the new starter's questions in context, with a link to the source document. What it cannot answer, it routes to the right person — and logs, so you learn what your documentation is missing.
What good looks like
New starters get consistent answers immediately instead of inconsistent answers eventually. Ramp-to-productive shortens, and the gap log quietly becomes the best documentation backlog the firm has ever had.
03 — The problem, in money
- New hires per month6 hires100 hiresPlaceholder — not a validated benchmarkStarting point
- Onboarding hours per hire8 hours41 hoursPlaceholder — not a validated benchmarkStarting point
- Share automated0.4 ratio0.5 ratioPlaceholder — not a validated benchmarkStarting point
- Blended blended_hourly_cost cost40 USD110 USDPlaceholder — not a validated benchmarkStarting point
- Share of freed time redeployed0.5 ratio0.5 ratioPlaceholder — not a validated benchmarkStarting point
04 — The options
05 — The call
Buy it. An off-the-shelf tool wins this one — it gets you live faster, and the math does not justify building from scratch.
06 — How this goes wrong
- Grounding it in documentation that is out of date. The assistant will confidently repeat a superseded policy, and new hires have no way to know.
- Including HR material with access rules — salary bands, performance notes, anything personal — in the same index as general policy.
- No escape hatch. A new hire stuck in a loop with a bot on day three learns that the firm does not want to be asked things.
- Measuring deflected questions instead of time-to-productive. Deflection is easy to game and is not the outcome you want.
07 — How you would actually do it
- 01Audit what documentation actually exists and how stale it is. This usually reveals the real problem.
- 02Scope the index to general policy and process; exclude anything with per-person access rules.
- 03Require a source link on every answer so the new hire can verify.
- 04Add a one-click handover to a named buddy.
- 05Log unanswered questions and work that list monthly.
- 06Measure time-to-first-independent-delivery, not chat volume.
08 — Tools worth looking at
Named for orientation, not endorsement. No affiliate arrangements, and nothing here has been paid for.
09 — Run it on your numbers
Start from industry averages, then drag the inputs to match your business. The number moves live.Drag each input to your number — the starting points are neutral midpoints, not industry data. The number moves live.
10 — The evidence · what others report
Reported numbers are third-party claims, source-attributed — not verified by me, and never blended into the model above.
11 — Questions owners ask
Is this your business?
Get your free AI Opportunity Score — the same first pass I run at the start of every Audit.
Get Your Free Score