The situation
Job applications arrived as unstructured emails: CVs, links, mixed languages, agency forwards, and colleague referrals. HR teams had to read every message, determine whether it was a real application, check eligibility, complete missing data, and register the candidate in the HR system.
At higher volumes, this created slow first responses, duplicate candidate risk, unnecessary back-and-forth, and compliance exposure around eligibility rules.
What we built
We rebuilt the process as a single AI-powered intake pipeline. Incoming emails are automatically classified as applications or non-applications, candidate and vacancy data are extracted, and policy rules are applied immediately.
From there, the pipeline handles the next step without manual input: requesting missing information, rejecting ineligible candidates in the correct language, detecting returning applicants, and creating valid candidates in the HR system including their CV.
The result
Requesting additional applicant info went from 7 minutes to 1 minute per application (86% faster), and syncing candidates to AFAS Profit from 13 minutes to 1 minute (92% faster). Policy checks run automatically on 90% of applications. Recruiters spend their time on qualified candidate conversations and hiring decisions instead of inbox administration.
As one HR Manager put it: everything is sorted automatically and the right next step is already initiated before anyone opens the inbox.
