Interview scheduling without the email ping-pong

The workflow reads the panel's calendars, applies your interview rules, offers the candidate a short list of real options, and books the loop with video links and interviewer briefs attached. Recruiting coordinators handle the exceptions instead of every thread.

01the problem and who owns it

A four-person onsite loop can take a dozen emails to arrange, and one reschedule restarts the puzzle. Coordinators spend their week reading calendars, candidates wait days for a slot, and good people accept other offers in the gap.

Recruiting ops owns the process, but the hard constraints come from interviewers: who can run which interview, who needs a break between sessions, which time zones are fair to the candidate. Those rules usually live in a coordinator's head.

02what the AI does, step by step

1. Read the stage and interview planWhen a candidate enters an interview stage, the workflow loads the plan from the ATS: which interviews, how long each runs, and which interviewers are qualified for each slot.
2. Search calendars under your rulesIt queries free and busy time for eligible interviewers and applies constraints: no back-to-back sessions for the same person, interviewer load limits per week, and hours that fall within the candidate's stated window and time zone.
3. Offer the candidate optionsThe candidate gets a link or email with a few complete loop options, not single slots. A model drafts a short, friendly message in the recruiter's voice that explains what each session covers.
4. Book everything at onceOn selection, the workflow creates calendar holds, video links, and room bookings, then writes the confirmed schedule back to the ATS.
5. Send interviewer briefsEach interviewer receives the resume, the focus area for their session, and the scorecard link, timed for the day before.
6. Handle reschedules and declinesIf an interviewer declines, the workflow finds a qualified substitute or a new time and asks the coordinator to approve the swap before the candidate is contacted.

03systems it connects to

• ATS. Greenhouse, Lever, or Ashby for interview plans, stages, and scorecards.
• Calendars. Google Calendar or Microsoft 365 through their calendar APIs.
• Video conferencing. Zoom, Google Meet, or Microsoft Teams for meeting links.
• Room booking. Whatever system manages conference rooms for onsite loops.

04human checkpoints

• Rules review. The recruiting lead approves interviewer eligibility and load limits before launch, since those encode fairness to interviewers.
• Substitutions. A coordinator approves any change to who interviews the candidate.
• Sensitive candidates. Executive searches and confidential roles are excluded from automation and scheduled by hand.

05what to measure

• Time from stage change to booked loop. Median and worst cases, per role.
• Reschedules per loop. Including who initiated them.
• Interviewer load spread. Whether the same few people are absorbing most interviews.
• Candidate drop-off before interview. Withdrawals and no-shows between scheduling and the first session.

06risks and guardrails

• Calendar data exposure. Reading calendars can surface private event titles. Request free and busy access only, never event details.
• Accommodation requests. Candidates who need more time, a different format, or an accessible location must be able to reach a person easily, and those requests are handled under ADA obligations, not by the scheduler.
• Tone failures. Generated messages to candidates should come from approved templates with limited variation, so nothing promises outcomes or reads as careless.

07build vs buy

Dedicated tools such as GoodTime or ModernLoop, and the schedulers built into Greenhouse and Ashby, handle standard loops well. If your process fits their model, buy.

Custom work is worth it when your loops have constraints those tools cannot express, such as clinical shift schedules, multi-site panels, or interviewer certifications held outside the ATS.

08related playbooks

Browse every hr and recruiting playbook or the full library.