Think of Megan like the teammate who never gets tired of resumes. Turn on AI Screening (and optionally AI Guidance) and she’ll evaluate applicants right when they hit your Applied stage; prioritizing the ones most likely to shine, and flagging the ones that need a closer look. You can toggle AI Screening and Guidance per job stage, and keep full control the whole way. (More on guardrails below!)
What Megan actually evaluates (beyond the basic matching)
Megan reads for depth, not just keywords. She looks across experience, skills, and demonstrated work to spot whether it’s real project experience or surface-level familiarity. She also weighs the overall balance; problem-solving creativity, technical rigor, and leadership—so the recommendation matches the real needs of the role (not just the phrasing of your JD). When data is ambiguous (e.g., unclear education), she can reach out to the candidate to clarify before making an early-stage decision.
You’ll also see helpful context like quick summaries and red flags (e.g., name/email mismatches), so you can spot issues fast without digging.
Where AI Screening runs
By default, AI Screening happens in the Applied stage of your job. That’s the ideal moment to let Megan do the heavy lifting; she screens every applicant and organizes outcomes so your team can focus on the best fits first.
Quick setup (2 minutes)
Open the job’s stages and select Applied.
Toggle on AI Screening. You’ll see three core options you can enable (plus optional guidance settings—see next section).
(Optional) Enable AI Guidance for extra signal-boosting checks.
(Optional) In the same stage, you can also enable AI Scheduling and AI Background Checks when you’re ready to streamline later steps.
Tip: You can still keep human review before any move: see “Controls & guardrails” below.
Add “Guidance” for the nuanced stuff (like education)
Want Megan to pay special attention to post-secondary education? Add guidance that highlights relevant programs/schools for the role and instructs Megan to ask the candidate when anything is unclear. She’ll contact them, resolve the ambiguity, and only then make a call for that early stage.
Teach Megan what you really care about
Job descriptions aren’t always the truth of what the team values. You can manipulate the requirements Megan uses—tighten them, relax them, or reweight them—so screening reflects the real bar for the role (not just the published JD).
And screening isn’t binary. If a candidate looks promising but a key requirement is missing from their resume, you can instruct Megan to reach out during screening to verify that capability, then continue the evaluation and update the record based on the response.
What you’ll see while Megan works
Real-time context & notes. Megan explains why she advanced or held a candidate, so you’re never guessing.
Tasks & timelines. Every action creates a task you can open to see each step she took—complete visibility across jobs.
Two-way sync with your ATS. Jobs, stages, candidates, interviews—everything stays mirrored and updated in both systems.
Controls, trust, and guardrails (you’re in charge)
Human-in-the-loop: If a question goes beyond her configured knowledge, Megan escalates to the hiring team with an @mention so a human decides.
Access you control: You add Megan as a Greenhouse teammate to specific jobs. Remove her, and related data access is revoked—the local copy is wiped on the next sync (within ~30 minutes). Add her back later and it resyncs.
Bias & compliance (in progress): We’re pursuing ongoing, region-aware AI certifications (including bias evaluation) as part of our program.
Example: a smooth early-stage flow
Candidate applies → lands in Applied → AI Screening runs.
Megan spots strong signals, but education is unclear → she messages the candidate to clarify.
Candidate confirms details → Megan updates the record and advances to shortlist with a short rationale in the timeline.
(Optional) From the same stage, AI Scheduling can kick off to coordinate interviews across your team’s calendars.
FAQs
Can Megan screen against “my version” of the requirements, not just the JD?
Yes: Edit the requirements she uses so they reflect the true bar.
Will she ping candidates to verify missing info?
If you enable it, yes—Megan can message the candidate during screening to validate capabilities and then continue.
What if I want to approve emails or stage moves first?
Keep human review on. Megan will draft, then wait for your go-ahead (and she escalates questions she can’t confidently answer).
How do I turn it off?
Open the job’s Applied stage and toggle off AI Screening (and AI Guidance if you were using it).
TL;DR
Flip on AI Screening at Applied, tailor the requirements, and add a bit of Guidance for the tricky bits. Megan will handle the first pass with transparency, escalate when needed, and keep everything synced—so your team spends time where it matters: real conversations with the best candidates.