Onboard a new hire
Day one for a new hire is rarely about the job. It’s a scavenger hunt — which Slack channel is the right one, where the deploy runbook lives, who to ask about payroll, what the password policy is. Every one of those questions costs a teammate a context switch, and the new person spends their first week feeling like a tax on everyone around them. Multiply that by the role being remote, and the handbook nobody has time to update, and the running start turns into a slow crawl.
Skynet flips it. You connect your handbook, wiki, HR docs, and the notes from the last few people you onboarded, and an agent answers from all of it — grounded, with the source cited, so the new hire trusts it and you can check it. The questions that used to land in a senior person’s DMs land with the agent instead. Your team stays heads-down; the new person gets unblocked in seconds.
How the onboarding agent works
Ground it in what you already wrote
Point the agent at your handbook, internal wiki, onboarding checklists, and past role-specific docs. Skynet builds unified memory from all of it, so when the new hire asks “how do I request time off?” the answer comes from your policy, with a link, not from a guess.
Generate a personalized 30-60-90
Give it the role and the person’s manager, and the agent drafts a 30-60-90 plan tailored to that job — the ramp goals, the people to meet, the systems to learn. The manager reviews it, trims a line, and shares it as the real plan instead of a blank template.
Build the access and intro checklist
The agent assembles the first-week checklist: the tools to provision, the channels to join, the recurring meetings, and the intros to make. Because Skynet takes action in the tools you use, it can draft those access requests and the welcome message for you to send.
Answer questions all week
For the rest of week one, the new hire just asks. “Where’s the design system?” “Who reviews infra PRs?” “What’s our on-call rotation?” The agent answers from your docs and cites where it found it — so they self-serve instead of queuing behind a busy teammate.
The whole thing is built from a plain-language prompt — no code, no flowchart. Describe the job once and reuse it for every new hire.
Where this lands
The new hire gets answers in seconds instead of waiting on a DM, a plan that fits their role, and a first week that feels like progress. Your team gets to keep working. And the onboarding knowledge stops living in one senior person’s head — it lives in an agent that every future hire can ask.
Frequently asked questions
The agent is only as current as what you connect — which is the honest version, because every answer cites its source. When a new hire hits a stale doc, you see exactly which one to fix, so onboarding becomes the thing that keeps your handbook honest instead of the thing that exposes it.
Only to the sources you connect, and only for answering. It drafts access requests and intro messages, but a human approves anything that provisions a tool or sends a message — you decide where its autonomy ends.
Yes. You describe the role and reporting manager in the prompt, and the agent tailors the 30-60-90, the checklist, and the intros to that job — an engineer and a salesperson get different plans from the same agent.
No. They ask questions in Chat the way they would ask a person, in plain language. The agent can also live in Slack, so onboarding happens in the channel they are already in on day one.