Case study · Recruiting
Recruit Garden — candidate screening, scored
80,000+
candidates enriched and scored — and counting, daily
~55
seconds per candidate, fully hands-off · no recruiter time
0–100
match score, with the reasoning written out factor by factor
Overview
The person you need for the role that landed this morning might already be in your database. Finding them means reading it.
That's the quiet tax on every recruiting agency. The candidates are an asset — sourced, interviewed, paid for — but they only pay out when someone opens them. And opening them means résumé, LinkedIn profile, recruiter's notes, holding skills and seniority in your head against a spec. A database of thousands is worth exactly as much as a recruiter can get through in a day.
Recruit Garden closed that gap inside the ATS they already use, without handing the decision to a model. The system enriches every candidate from LinkedIn, résumé and recruiter notes, scores them against a specific role, and writes back a verdict a human can check and overrule. 80,000 candidates in, it runs every day.
- Client
Recruit Garden
Before
The database was only as useful as the last person who read it.
Screening ate the most expensive hours in the business and returned the least. Every shortlist meant a recruiter working down profiles by hand, weighing skills and seniority against the spec in their head. Two recruiters could look at the same candidate and rank them differently — and neither could tell you afterwards why. The judgment lived in the reading and left with the reader.
Left unchanged, it kept skilled recruiters buried in profiles instead of talking to the best people. And the candidates already in the system — people the agency had sourced, interviewed and paid to acquire — went unseen, because nobody had time to re-read thousands of profiles every time a new role came in.
The judgment lived in the reading and left with the reader.
After
A database that keeps itself current — and a shortlist in one click.
Two things, in that order. The scoring is what recruiters asked for. The data underneath it is what made the scoring worth trusting.
First, the records fix themselves.
Before a candidate is ever scored, the system reads their live LinkedIn and their résumé and refreshes the ATS profile with what's actually true now — current seniority, skill set, company, location. Stale profiles get re-pulled automatically. Records that used to start decaying the day they were saved now keep themselves current. That alone turned a dormant candidate list into a clean, searchable asset in its own right — and it's the ground every score is calculated on.
Then, the shortlist.
A recruiter opens a role and presses one button. Every candidate comes back scored 0–100 against that specific role, with a verdict — Strong / Potential / Filtered — and the reasoning spelled out skill by skill. Instead of reading everything, they review a ranked list and can see exactly why anyone landed where they did. When they disagree, the reasoning is right there to argue with. The decision still belongs to the recruiter.
Built so the score can be trusted.
The number isn't the model's opinion. Scoring is deterministic maths over extracted factors, so the same candidate against the same role scores the same way every time. The model's job is deliberately narrower: read the sources and quote only evidence that's actually present. It can't credit a candidate with a skill nobody wrote down.
The outcome
The database becomes the advantage.
Most agencies compete on who sources fastest. Recruit Garden can compete on who already knows the right person — because the candidates they have spent years collecting are current, comparable, and one click from a shortlist.
And it compounds: every candidate added from here is scored against every role that follows, with no extra effort from anyone.
80,000+
candidates enriched and scored — and counting, daily
We're living our best life since your automation went live. We've already updated 8,000 candidates — and we're doing it every single day.
Sitting on a database nobody has time to read?
Book a 20-minute call. We'll show you exactly how we would work for your team — no slide deck, just the live system.
Speak with our team
Twenty minutes, no slide deck.
Tell us what's eating the week. We'll show you the live system and what we'd wire first.
Got it — we'll be in touch.
Expect a reply within one working day, with two or three times that suit us.