Client & Candidate Communication: Candidate Outreach Sequencer

Prompt: Candidate Outreach Sequencer

Description: Generates personalized first-touch outreach for a list of target candidates on a specific search.
Frequency: Ad hoc (at search kickoff or pipeline refresh)
Connectors: Clockwork Recruiting, Gmail / Outlook (or Slack for internal handoff)

Prompt: (copy/paste the prompt below into your chat in your LLM)

Project: [project name]. Fetch with ?include=client_company and display as <Company> \ <Project>.

Call list_reference_data(data_type=statuses) to confirm the categories for "not yet contacted" candidates (research-researching and early in-outreach). Pull the project brief and the candidacies in those categories via list_candidacies with include=person,status.

Filter out status.rank < 100.

Before drafting outreach for anyone: run run_conflict_check(person_id=...) and check doNotContact on the person record. Exclude anyone returning STOP or flagged DNC, and mark them "Excluded — [reason]". Surface CAUTION results for my review rather than auto-excluding.

For each remaining person, pull person details and prior notes, then draft a personalized outreach email that:

  • Opens with a specific, non-generic hook from their background or a prior Clockwork interaction
  • References the role in 1 sentence without naming the client unless the project is non-confidential
  • Asks for a 20-minute exploratory call with two proposed time windows
  • Avoids recruiter clichés ("rockstar", "ninja", "exciting opportunity")

Output: a table — Name | Current Company | Hook | Draft Email | Conflict/DNC Flag. Flag anyone I've personally interacted with before for review.

Adapt It: Generate LinkedIn/InMail variants instead of email Produce a 3-touch sequence (intro, follow-up, break-up) per candidate Adjust seniority/tone; batch by company for a coordinated approach Route the drafts to a researcher via Slack for sending

Did this answer your question? Thanks for the feedback There was a problem submitting your feedback. Please try again later.