Skip to content
SeekOutDevelopers

Guides

Make your next hire possible.

Turn a role into a useful shortlist, a market brief, or a workspace your team can use. Start with the tools your company already has.

Find my starting pointShare with an agent

Use an existing assistant, or have an agent build something for your team.

Start where your team already works

  • Microsoft 365 Copilot
  • ChatGPT
  • Claude
  • Gemini Enterprise
  • Amazon Quick
  • Databricks

What would help you today?

Common boundaries

Build a hiring-manager shortlist

A ranked slate with evidence, gaps, and suggested interview focus.

  1. Clarify must-haves, nice-to-haves, title variants, locations, and target companies.

  2. Select accessible data sources and run preview searches before showing candidates.

  3. Rank by evidence and name uncertainty separately from confirmed profile data.

What the assistant should return

  • Candidate list includes match evidence, not just names.
  • Ranking rationale is tied to the role criteria.
  • Any paid or write action is paused for confirmation.
  • If the pool is too broad, tighten seniority, title, skill, or location constraints.
  • If the pool is too narrow, preserve must-haves and relax lower-priority filters.
  • If a source is unavailable, say which license, tenant, or permission is missing.
  • Search and comparison are read-only; contacts, exports, saves, and ATS pushes have separate credit or write effects.
Read the starting prompt
Build a shortlist of the five strongest candidates for our senior platform engineer role. Compare AWS, Kubernetes, and infrastructure-as-code evidence and explain the ranking.

Use the canonical Candidate search referenceFetch this recipe as JSON

Kick off with talent-market context

A role strategy with market depth, employer patterns, and constraints to relax.

  1. Start before a slate, with role level, core skills, locations, and tradeoffs.

  2. Compare markets from aggregate counts and facets before asking for profiles.

  3. Name requirements that narrow the pool and propose bounded relaxations.

What the assistant should return

  • The answer shows market depth and the constraints driving it.
  • Location and skill tradeoffs are separated from recommendation language.
  • No candidate data is invented when an aggregate market cannot be resolved.
  • If a market cannot be resolved, report it as unavailable instead of estimating.
  • If counts are too small, broaden one constraint and preserve explicit anchors.
  • If a vertical is not entitled, state the missing entitlement and continue only with accessible sources.
  • Market and preview search are read-only unless the workflow asks for paid facets or full-profile hydration.
Read the starting prompt
Compare the available pool of staff machine learning engineers in Boston, Seattle, and Austin. Show common employers and skills, and identify which requirements are limiting the pool.

Use the canonical Candidate search referenceFetch this recipe as JSON

Rediscover authorized ATS candidates

A ranked list of previous applicants or pipeline candidates worth revisiting.

  1. Confirm that ATS pipeline search is available before using applicant data.

  2. Search by role, stage, date range, disposition, notes, or accessible fields.

  3. Rank rediscovery candidates by current fit and prior-stage signal.

What the assistant should return

  • The answer names the ATS source as connected and authorized.
  • Unavailable ATS access is explicit rather than replaced by public search.
  • Ranking separates past pipeline signal from current profile evidence.
  • If ATS is not connected, tell the user to configure it before rediscovery.
  • If the user lacks access, stop and do not infer private pipeline data.
  • If a requisition is ambiguous, ask for the destination before writing or exporting.
  • ATS search can be read-only; exporting or pushing candidates is a write and may use credits or destination-specific permissions.
Read the starting prompt
Search our ATS for past candidates who could fit a Senior Backend Engineer role. Prioritize later-stage candidates and explain why each one is worth revisiting.

Use the canonical Candidate search referenceFetch this recipe as JSON