‹ Recruiting & Onboarding recipesRecruiting & Onboarding · Agent Blueprint
Pre-loop sync
Candidate dossier for the next interview loop. Everything the team has ever said about this candidate, surfaced in one place.
Side by side: token usage, with and without Paperbase
~94k tokens saved · 78% less work for the agentWithout Paperbase
~120k tokens
- Search email for the candidate's name and pull every recruiter thread.
- Skim the ATS notes per interviewer.
- Search Slack for mentions of the candidate.
- Read each interviewer's comments and infer their stance.
- Reassemble the picture from fragments and memory.
With Paperbase
~26k tokens
- Pull every Interaction involving the candidate across recruiter, interviewers, and team.
- Ask Paperbase for each participant's extracted Stance.
- Surface disagreements between interviewers automatically.
- Produce the dossier, quoted and sourced.
Agent prompt
You are preparing a candidate dossier for {{candidate_name}} before their loop.
Using Paperbase:
1. Pull every Interaction mentioning {{candidate_name}} - recruiter threads, interview notes, internal Slack.
2. For each reviewer, extract their stance: signal, concern, or hold.
3. Flag where reviewers disagree, quoting each side.
4. Note anything unresolved from prior rounds.
Output:
- The candidate, in one paragraph, from the whole team's record.
- Per-reviewer stance, quoted.
- Disagreements the loop should resolve.
- Questions to carry into the next interview.Placeholders in {double_braces} are inputs the agent will ask for at runtime. Give the prompt to any agent connected to Paperbase, and the rest grounds in your own memory graph.