Some recruiting tasks are worth automating partway. This one you can hand over almost entirely, and you probably should.
Interview scheduling is the back-and-forth that nobody went into recruiting to do: matching a candidate's availability to a panel's calendars, sending the invite, fixing it when someone moves a meeting. The case for automating it isn't that it eats your week. The hours it saves are real but modest against a hiring process measured in weeks. The case is simpler: it's mechanical, there's no judgment call in it, so it's the one task you can hand off completely, and doing so is easy. There's no good reason to keep spending human time on it. Speed helps too: a candidate booked in minutes instead of days is less likely to cool off or take another offer first.
The good news is that this is exactly the kind of task AI is best at, and the tools have caught up. Self-scheduling turns a multi-day, multi-email booking process into a single click for the candidate, and a coordination agent handles the calendar math and the reschedules on its own. It automates this cleanly for a structural reason: scheduling is logistics, not a decision about the candidate. No one gets screened out by a calendar.
This guide walks the three rungs of automating it, with a concrete setup, the tools that actually do this in 2026, and the one situation where you should still pick up the phone yourself.
The job to be done
Scheduling sits between two decisions you've already made. You've decided this candidate advances, and you know who's on the panel. The task is purely operational: find a slot that works for everyone, confirm it, and keep it confirmed when life happens. Three sub-tasks, all mechanical. Offer times that fit real availability. Book the slot and send invites with the right links and details. Handle changes when a candidate or interviewer reschedules.
None of this involves judging the candidate, which is the whole point. It's worth being precise here, because two different things get called "AI interviewing." Scheduling an interview is logistics and carries no legal weight. Conducting or evaluating an interview with AI is a decision about a person, and that one is regulated. This guide is about the first. Keep them separate in your head, because the automation advice is opposite: automate scheduling all the way, and keep a human firmly on evaluation.
The automation ladder
Every recruiting task can be automated to a different degree. We use three rungs.
| Rung | What it means | Effort |
|---|---|---|
| L1: one-shot prompt | You ask a chat tool to draft the scheduling email or work out a slot across timezones. You still send and track it. | Minutes |
| L2: saved playbook | A booking link plus reminder templates. The candidate self-schedules into your real availability, and you stop writing the emails. | An hour to set up |
| L3: full agent | A scheduler syncs every calendar and your ATS, books the panel, sends reminders, and reschedules on its own. You only touch exceptions. | Tool config |
Unlike resume screening, which stops at L2, scheduling reaches L3 and stays there. The mechanical nature of the task is precisely what lets it run unattended.
L1: the one-shot prompt
The entry point, free, no setup.
You paste the constraints into a general assistant like ChatGPT or Claude: "The candidate is in London and free Tuesday and Wednesday afternoons. The panel is in New York and Austin. Propose three one-hour slots that work for all three timezones next week, and draft the invite email." You get back usable options and a clean draft in seconds. For a single interview with simple constraints, this beats doing the timezone arithmetic yourself.
Where it breaks: the model can't see anyone's real calendar. It's guessing at availability from what you typed, so every slot it proposes still needs you to check it against actual calendars and send it manually. L1 saves you the writing, not the coordination. It's a drafting aid for one interview, not a system.
L2: the saved playbook
This is the rung that pays for itself, and where most recruiters should start.
Instead of proposing times, you send a booking link tied to your real availability. The candidate picks a slot, it lands on the calendar, and reminders go out automatically. You don't have to pay for this. Google Calendar now has appointment scheduling built in, so a free personal account can publish one booking page at no cost (paid Workspace plans add reminders and unlimited pages). Calendly is the familiar standalone tool, and cal.com is the open-source one, with a genuinely capable free tier for individuals and the option to self-host if you want to own the data. If you book more than one kind of interview, a 15-minute phone screen and a 60-minute panel, or a separate link per client, that means multiple event types: cal.com includes them on its free tier, Calendly puts multiple event types and multi-calendar conflict checking on a paid plan, and Google's free page is a single type. For hiring specifically, the better setup pulls from a panel's combined availability rather than just yours. Either way you've removed the entire email thread: the candidate schedules themselves, around your real constraints, in one click.
The payoff shows up fastest in candidate experience. A booking link collapses that multi-day scheduling delay into a single click, and faster scheduling is one of the clearest ways to stop losing candidates between stages. For high-volume roles, this single change is the highest-return automation on the whole desk.
Where it breaks: complex panels. A four-person loop with rooms, sequencing, and interviewer load-balancing is more than a booking link can hold. That's the gap L3 fills.
L3: the full agent, and why scheduling actually gets there
L3 is a system that owns the whole coordination job. It reads every calendar and your ATS, finds slots that satisfy the candidate and the full panel, books rooms, sends the invites and reminders, and rebooks automatically when someone drops. For a multi-stage loop, this is the difference between an afternoon of coordination and a few minutes of review.
This is where scheduling separates from almost every other task on the recruiting map. Screening stops at L2 because the judgment and the law both push back. Scheduling has neither problem. There's no protected decision being made and nothing to audit, so the autonomous version is genuinely usable today, and the tools below do it in production.
| Tool | Best for | What it does |
|---|---|---|
| GoodTime | Enterprise, complex multi-panel loops | Its Orchestra system assigns interviewers by training status, swaps panelists when conflicts arise, detects bottlenecks, and runs the workflow autonomously. |
| ModernLoop | Fast-growing in-house teams | Syncs calendars and ATS to find slots automatically, supports panel and group interviews, and reschedules when plans change. |
| Paradox (Olivia) | High-volume frontline hiring | Conversational AI texts candidates directly and books the interview inside the chat, with no recruiter in the thread. |
| Calendly, cal.com, Google Calendar | Solo recruiters, simple loops | The L2 booking-link layer rather than full L3 coordination. cal.com is open-source and free (multiple event types, calendar sync, self-hostable); Calendly's free plan is one event type and one calendar, with per-persona links and up to six calendars on paid; Google Calendar gives one free booking page. |
What still needs a human
Two things, and neither is the coordination itself.
The first is the exception. Senior and executive candidates, sensitive reschedules, a finalist you can't afford to lose: these are moments where a personal email or call does real work, and routing them through a bot reads as cold. The skill at L3 is not doing the scheduling. It's knowing which 5% of interviews to pull out of the automation and handle yourself.
The second is the decision that comes before scheduling, which is who gets the interview at all. That's screening, and it carries the judgment and the legal weight that scheduling doesn't. Automate the calendar completely. Keep yourself on the call about people.
Where Glozo fits, and where it doesn't
Straight answer: Glozo doesn't schedule interviews, and it won't pretend to. For the task in this article, the right tools are the ones above. Pick one and turn it on.
Where Glozo plays is earlier in the same week. Scheduling load is downstream of sourcing volume, and Glozo works at the top: its Smart Search reads intent rather than keywords, surfaces candidates from 30+ sources by how closely they match the role, and flags who's actually open to a move, so the people who reach the scheduling step are worth the panel's time. Fill the funnel with better-matched candidates and every interview you book is a better use of the hour. That's an honest division of labor, not a pitch to replace your scheduler. (For the ATS side of the same workflow, our guide to open-source ATS tools covers the systems these schedulers sync with.)