Recruitment: Open Req to Hire

A recruiter creates the requisition from the Job Reqs list: title, employment type, location, remote policy, and the salary band. The system seeds the same five-stage pipeline every role uses — Screen, Interview, Offer, Hired, Rejected — so no two roles ever have subtly different processes. The req opens for hiring immediately but stays off the public site until the posting is ready.

The public posting is written in markdown on the req itself: the job description, a clean URL slug, and up to three screening questions every applicant answers. Flipping it public puts the role on the careers page; closing the req later takes it down automatically. Internal notes stay internal — the posting and the working record are two views of one req.

Door one: the careers page. A candidate applies with a resume, cover letter, and answers to the screening questions. The application lands in the talent system attached to the req — no inbox, no forwarding.

Door two: the recruiter batch drop. Sourced candidates — from LinkedIn, referrals, or agencies — enter through the same funnel. A recruiter drops up to 25 resumes at once, and AI reads each file and prefills a draft: name, email, phone, LinkedIn, headline, current title. The recruiter reviews each draft and saves it. Ten minutes of drag-and-drop replaces an afternoon of data entry, and both doors produce identical records.

Duplicates cannot happen by construction: candidates are keyed by email, and one person holds at most one application per role. Re-adding someone surfaces their existing application instead of silently creating a second.

The moment an application lands — from either door — Claude reads the full resume against the job description and writes a structured screen: a 0–5 fit rating, an overview paragraph, the candidate's concrete strengths and gaps, an English-proficiency read, and the salary expectation and notice period exactly as stated (never guessed).

That assessment travels with the candidate everywhere: it appears in the application's side panel, on the req's ranked list, and in the candidate pool. Every application gets the same depth of read whether it arrived first or five hundredth — screening quality no longer depends on who reads the pile.

Scored candidates also stack-rank within their role family, not just the single job they applied to. A strong engineer who applied to the wrong opening still surfaces near the top of the engineering family instead of dying in the wrong folder.

Each req has a kanban board: candidates move Screen → Interview → Offer by drag, and every application carries two ratings side by side — the AI's score with its written reasoning and the recruiter's own star rating. The list sorts by either one. Candidates strong on both gates make the shortlist; where the gates disagree is exactly where a second human look goes, instead of a quiet rejection.

All working context lives on the application: a notes thread, the resume (replaceable if a better version arrives), the cover letter and question answers, and the candidate's profile. Anyone on the team can open any application and know exactly where it stands.

From the shortlist on, the process is entirely human: interviews, references, and the offer. AI gets every resume read; people decide who joins. A candidate leaves the pipeline with an explicit status and a recorded reason — hired, rejected with a note, withdrawn, or parked as future consideration. Nobody is rejected by AI alone, and nobody is left in limbo.

A hire flips the application status and hands off to the New Member Onboarding workflow, which turns the applicant record into an employee record without re-typing anything.

Every person who has ever applied — hired, rejected, or parked — stays in one searchable, sortable pool, stack-ranked by their best AI screen and grouped into role families. When the next role opens, sourcing starts from everyone the company has already met, not from zero.

The strong runner-up from last quarter's search surfaces at the top of the next one, with the full history attached: every application, every screen, every note.