How We Saved a UK Recruitment Agency 25 Hours Per Week with AI Candidate Screening
LookADev engineered an autonomous semantic candidate screening agent integrated with Bullhorn ATS for a UK recruitment agency, cutting CV screening time by 80%, saving 25 hours per week (£36,000/year in senior recruiter capacity), and achieving a 45-day financial payback.
3 senior recruiters spent over 25 hours per week manually opening, reviewing, and scoring 300+ applicant CVs inside Bullhorn ATS, delaying client shortlists and losing placements to competitors.
£36,000/year in billable senior recruiter hours squandered on manual CV review instead of high-margin client placement.
Absence of a semantic parsing layer inside Bullhorn ATS: legacy keyword searches either excluded top non-standard profiles or flooded inboxes with unqualified applications.
Autonomous semantic screening agent built with Next.js 15, PostgreSQL pgvector, and Bullhorn REST API. Evaluates incoming CVs against job specifications, calculates match confidence, and syncs structured recruiter scorecards in 8 seconds.
Deterministic candidate scoring and ingestion pipeline ranking top 5 applicant matches instantly and syncing notes back to Bullhorn with zero manual copy-pasting.
45-day project payback (£36,000/yr in senior recruiter capacity recaptured) plus a 30% acceleration in client shortlist submissions.
80% reduction in screening time, sub-8s candidate evaluation turnaround, 25 hours saved per week, and a complete financial payback in 45 days.
Facing a similar operational bottleneck in your business?
Schedule an AI Systems Audit (48h) with Lucas Martins. We diagnose exactly where your workflows are leaking time, calculate your ROI before writing code, and map a fixed-price production sprint.