Hiring docs are scattered and sensitive
When a startup grows from five people to fifteen, the founder suddenly needs to produce a lot of hiring documentation. Role descriptions for three open positions. Interview scorecards so the team evaluates candidates consistently. A team structure doc that maps out who reports to whom and what the org looks like in six months.
Most founders draft these in Google Docs or Notion, then paste pieces into ChatGPT to clean up the language. The problem is obvious once you think about it. Your hiring plan contains salary ranges, equity allocations, and notes about current team members' performance. You're sending all of that through a cloud service.
And the AI has no context. It doesn't know what your other role descriptions look like. It can't check whether the senior engineer requirements conflict with the junior engineer ones. Every role description gets written in isolation.
One folder, all your hiring docs
In Ritemark, you create a hiring folder. Inside it goes everything: the team structure overview, each role description as its own file, the interview scorecard template, and your notes about compensation bands. The AI agent can read all of these files when you're working on any one of them.
You open the product designer role description and ask the agent: "Read the senior engineer role description and make sure the collaboration expectations are consistent between the two roles." The agent checks both files and flags a mismatch. The engineer description says "works independently with minimal oversight" while the designer description expects "close daily collaboration with engineering." You fix it before posting either role.
The agent also catches gaps. "This role description doesn't mention anything about the reporting structure. Based on the team structure doc, this person would report to the CTO. Should I add that?" Small things, but they make a difference when a candidate reads the listing.
Building scorecards that match the roles
Interview scorecards are tedious to write from scratch. But they're important because without them, each interviewer evaluates candidates on different criteria and the debrief becomes a mess of gut feelings.
You draft the scorecard template once. Then for each new role, you ask the agent: "Based on the product designer role description, generate an interview scorecard that covers the key requirements. Use the same format as the engineering scorecard." The agent reads both files, maps the role requirements to evaluation criteria, and produces a scorecard that's consistent with your existing ones.
You adjust the weights, add a question you always ask, and save it. When you hire the next role, the agent already knows your scorecard format and your evaluation style. The documents build on each other.
Sensitive by default
Hiring plans are some of the most sensitive documents a founder writes. Salary bands reveal your budget constraints. Org structure notes might mention that a current role is being replaced. Performance observations about existing team members might inform new role requirements.
All of these stay on your machine. The AI agent reads them locally. No cloud storage, no third-party indexing, no accidental sharing. When you close the folder, the documents are exactly where you left them and nowhere else.