A Senior Product Manager role. 5 shortlisted candidates with complex constraints — a Dubai timezone, a single two-hour availability window, a mid-sprint engineer, and a reschedule. Our AI scheduling pipeline handled all of it, automatically, in under four minutes. Here's exactly what it produced.
"This is a live demonstration using our production scheduling pipeline. These are simulated candidates but the AI processing, calendar logic, and email drafting are real."
5
Candidates coordinated
3m 12s
Fully confirmed
21
Emails generated
~60s
Recruiter time needed
It's not one meeting. It's fifteen small interruptions spread across three days.
| Task | Per candidate | 5 candidates |
|---|---|---|
| Check hiring manager calendar | 5 min | 25 min |
| Draft and send proposal email | 5–7 min | 25–35 min |
| Process reply, re-check calendar | 3–5 min | 15–25 min |
| Book invite + send confirmation | 3–5 min | 15–25 min |
| Total | ~15–20 min | 75–110 min |
That's nearly two hours — spread across three days of interruptions
Every email reply breaks focus. Every reschedule re-opens the thread. And when a candidate asks to swap slots, you're back to square one — manually checking a calendar you've already checked twice. The cognitive overhead of tracking five concurrent scheduling threads is real, and it compounds when you're managing multiple roles simultaneously.
From calendar analysis to confirmed interviews — fully automated, with one human approval checkpoint.
Candidates marked Ready to Schedule
Pipeline triggered automatically — no manual handoff
Hiring manager's calendar read via API
All 23 existing commitments mapped; available windows identified with 15-minute buffers
Each candidate's constraints analysed
Timezones, hard windows, employment constraints, and preferences all factored in
Constraint solver assigns optimal slots
Non-conflicting schedule produced in milliseconds — hard constraints prioritised first
Personalised proposal email drafted for each candidate
Each email references their specific profile, constraints, and context — not a template
Single Slack notification sent to consultant
Full schedule overview + draft summaries — one approval checkpoint, then done
Replies monitored and processed automatically
Acceptances → calendar invite + Meet link + confirmation email. Reschedules → re-analysis + new proposal.
Reminders sent automatically
24 hours and 1 hour before each interview — every time, without fail
| Metric | Result |
|---|---|
| Candidates coordinated | 5 |
| Time to all emails sent | 1 min 19 sec |
| Time to fully confirmed | 3 min 12 sec |
| Scheduling conflicts | 0 |
| Reschedules handled automatically | 1 (resolved in 38 min) |
| Calendar invites created | 5 (with Google Meet links) |
| Total emails generated | 21 |
| Consultant approval checkpoints | 1 |
| Consultant time required | ~60 seconds |
Zero conflicts. All five candidates slotted into the hiring manager's available windows despite 23 existing calendar commitments, a Thursday Amsterdam flight, a Dubai timezone, and one mid-sprint engineer with a single two-hour window all week.
| Step | Manual Process | ShortlistOps |
|---|---|---|
| Read hiring manager's calendar | 5–10 minutes | 1.2 seconds |
| Analyse 5 candidates' constraints | 15–20 minutes | 0.3 seconds |
| Find non-conflicting schedule | 10–15 minutes | 0.3 seconds |
| Draft 5 personalised emails | 25–35 minutes | ~8 seconds |
| Consultant review and approval | — | 60 seconds |
| Process replies and confirm bookings | 15–25 min (multiple sittings) | Automatic |
| Handle 1 reschedule | 10–15 min + extra emails | Automatic |
| Create 5 calendar invites with Meet links | 10–15 minutes | Automatic |
| Schedule 10 reminder emails | 5–10 min (if remembered) | Automatic, always |
| Double-booking risk | High (manual tracking) | Zero (agent enforces) |
| Total recruiter time | 110–155 minutes | ~60 seconds |
Three of the 21 emails the pipeline produced — each personalised to the candidate's specific situation.
Sample 1 — Candidate with a single available window
Hi James,
I've had a good look at Marcus's calendar alongside the window you gave me, and I'm pleased to say we've got a clean fit:
Wednesday 25th February at 13:30 GMT — 45-minute video call (Google Meet)
That's the slot you mentioned as your only available window this week, so I've gone ahead and held it exclusively for you. No conflicts on Kestrel's side, and it gives you a proper lunch break beforehand so you're not jumping straight out of a GoCardless sprint into an interview.
Marcus Webb is a very straight-talking interviewer — from what I've gathered, he's less interested in structured STAR answers and more in how you actually think through product problems. Given your work on GoCardless's payouts infrastructure and the merchant-facing product decisions you've been driving, you'll have plenty of material that's directly relevant.
I'll send the Google Meet link as soon as you confirm below.
AI-generated · Personalised from candidate profile · Sent in under 10 seconds
What makes this personalised: The email acknowledges James's hard constraint directly ("the slot you mentioned as your only available window"), references his employer and current project context (GoCardless sprint), and offers specific insight about the interviewer's style. A template cannot do this.
Sample 2 — Candidate in a different timezone (Dubai, GMT+4)
Hi Leila,
Lovely to speak with you last week — and I can tell you're genuinely excited about this one. The feeling is mutual on Kestrel's side, which is why I want to get something in the diary while the momentum is there.
Given you're in Dubai for the next few weeks, I've found a slot that works before your working day kicks off properly:
Tuesday 24th February at 09:00 GMT (13:00 Gulf Standard Time) — 45-minute video call
That gives you a comfortable morning before the Tabby day really starts, and sits within Marcus's earliest available window on his end. I'd suggest hopping on 5 minutes early just to test audio/video — timezone interviews have a way of surfacing tech gremlins.
Both timezones shown · Current employer referenced · Practical timezone advice included
Sample 3 — Automatic reschedule handling (Rachel Byrne)
Hi Rachel,
No need to apologise at all — these things happen, especially when you're in the thick of a busy process. I completely understand.
I've had a look at the rest of the week and Friday morning is our best option on Kestrel's side:
Friday 27th February at 11:00 GMT — 45-minute video call
I should be transparent: Marcus is actually travelling to Amsterdam on Thursday evening, so Friday would be a video call with him joining from there — but he's confirmed he's completely happy to do that, and it won't affect the format of the conversation at all.
If Friday doesn't suit, I can also offer the following week — Monday 2nd March would be the next clean option, and there's more flexibility in Marcus's diary from that point.
Reschedule parsed automatically · Re-analysis complete · Alternative proposed · No consultant involvement
What this email does well: It surfaces a genuine piece of context (Marcus is in Amsterdam) rather than hiding it. It offers a clear alternative. The whole reschedule cycle — reply received, calendar re-analysed, new proposal drafted and sent — completed in 38 minutes without a single consultant intervention.
For a recruitment agency scheduling 30 interviews per week.
8.5h
saved every week
vs. manual scheduling
32h
saved every month
per consultant
£13,400
saved annually
at £35/hr consultant cost
| Metric | Manual | ShortlistOps | Saving |
|---|---|---|---|
| Time per interview (full cycle) | 15–20 min | 1–2 min | ~17 min |
| Interviews per week | 30 | 30 | — |
| Time per week | 7.5–10 hrs | 30–60 min | ~8.5 hrs/week |
| Time per month | 30–40 hrs | 2–4 hrs | ~32 hrs/month |
| Cost (at £35/hr) | Manual | ShortlistOps |
|---|---|---|
| Monthly hours on scheduling | 35 hrs | 3 hrs |
| Monthly cost | £1,225 | £105 |
| Monthly saving | £1,120 / month | |
| Annual saving | £13,400 / year | |
And that's before accounting for the compounding effects.
Role: Realistic Senior Product Manager specification for a fictional London fintech (Kestrel Financial Technologies)
Candidates: 5 fictional candidates designed to represent genuine scheduling complexity — timezone differences, hard constraints, employed candidates, reschedule scenarios
Hiring manager's calendar: Realistic busy-week calendar with 23 existing commitments and an Amsterdam trip
Scheduling logic: Real constraint-satisfaction algorithm — the slot assignments shown are what the system actually computed
Emails: Generated by Claude (Anthropic's AI), personalised to each candidate's profile — exactly as the production system generates them
The candidates are simulated. The complexity is real. The AI processing is real. The time saving is real.
Book a demo and we'll run a live scheduling round with your actual candidates and your hiring manager's real calendar — on the call.
Or email us at anas@shortlistops.co.uk — we'll get back to you within 24 hours.
Call or WhatsApp: +44 7388 281312