From brief to justified shortlist
What used to take a couple of days of manual searching became a search in minutes. The recruiter spent the time assessing instead of looking.
Case study
SEVR activated its talent base with Week 29.
SEVR has a mature recruitment flow and a talent base worth activating. Week 29 was placed on top of their existing ATS, without replacing anything, and made the candidate data searchable by meaning. The result showed up in everyday work immediately.
Pilot resultsSEVR
previously about 2 days< 1 hrTime to the first candidate list with explanations
previously under half100%Applicants who got a reply
profiles that never surfaced before1 in 4On the list were previously invisible candidates
up to80%Less time searching for candidates
The starting point
Like most established recruitment teams, SEVR sat on years of candidates in its ATS. Structured fields, free text and interview notes. The value was there, but it could not be reached. The best profiles were often the ones that were never tagged correctly, and time went to searching instead of assessing.
What we did
01
Activated the talent base
Week 29 indexed SEVR's existing candidate data and made it searchable by meaning. CVs and free text were pseudonymised before AI. No operational flow was moved.
02
Ranking with a rationale
Matches were ranked against the requirement profile, with a rationale per placement that the recruiter could show the candidate and audit.
03
Feedback to everyone
Past applicants received tailored feedback, always after the recruiter's approval. Nothing was sent in silence.
What SEVR got out of it
No applicant left without a reply
Everyone who applied got feedback. It protected SEVR's brand and kept doors open for future roles.
Hidden candidates became visible
About a quarter of the shortlist came from profiles that earlier searches never caught. Value that was already in the base, but out of reach.
Less time searching for candidates
Most of the time that used to go to finding candidates was freed up for interviews and assessment.
- 4 in 5 top matches were judged relevant by the recruiter.
- The human adjusted the AI's order in about 1 in 5 cases, a sign of meaningful oversight.
- All processing happened within the EU, with pseudonymisation before AI and the decision with the human.
Week 29 made our existing talent base searchable in a way we did not think was possible. We find the right people in minutes, and no applicant is left without a reply.
How we measured
The numbers were measured in SEVR's own flow during the pilot, against their own baseline.
- Time to shortlist was measured from brief to the first justified list.
- Share of applicants with a reply was compared to SEVR's state before the pilot.
- Everything within the EU, pseudonymisation before AI, the decision with the human.
Your data, our engine
Want to see it on your own data?
We run the same search live on one of your real roles.
30 minutes in your own flow. You pick a real role, we run the search live.
Need to get internal buy-in?