Short answer
Let a system search your whole historical database on every new vacancy or request and surface the strongest candidates with reasoning. That is usually the cheapest source of candidates you already own.
Why that database sits still
Not because the people in it stopped being interesting. Searching an ATS is work, and on a new vacancy calling is faster than digging. So nobody digs, and the database fills up with people you once spoke to and then lost sight of.
How an activation works
On every new vacancy or request the system scans your whole historical database. The top 20 arrives in your inbox, with short reasoning per match: why this person, which experience fits and when you last spoke.
If you run a flexible pool, it fishes out your available and dormant temporary workers the same way.
Under the hood sits our own orchestration code, with Claude for the language work, Pinecone for the search and your ATS API for the rest. The result lands where your team already looks: the inbox or a Slack channel of your choice.
What you need to supply
- Access to the ATS API or the staffing software the database sits in.
- A picture of which fields your team really fills in and which stayed empty.
- Someone who reviews the first digests.
A first data cleanup is in scope when we activate your database. Messy data is not a reason to wait.
What stays human work
Calling. The system makes sure the right twenty names sit on top, your recruiter does the rest.
And the data itself stays in your own systems. We build a system around it, we do not take it out.