From 30 outreaches a day to the entire candidate pool
How BPM Company used AI coworker Lara to scale recruitment, from a manual bottleneck to a full screening pipeline of 24,298 candidates in a single quarter.
Measurement period
Jan – Mar 2026 (Q1)
24,298 candidates
0
Candidates screened
0
Personal InMails
0 sec
Per candidate
0
Per day (was: ~30)
We appreciated how AgentsLab involved us throughout the whole journey. It felt like we were building something together, rather than buying a black box.
Eelco Vissinga
CEO · BPM Company
01, Situation
The bottleneck was screening, not the market
Recruiters at BPM Company simply couldn't escape it: read profiles, compare against open roles, write a personal InMail, send. Per recruiter that added up to a full day for at most ~30 suitable outreaches.
The problem wasn't supply on the market, that's abundant. The problem was the human capacity to screen that supply thoroughly and consistently. Strong matches went undiscovered, and the focus stayed on outreach quantity instead of precision.
~30
Suitable outreaches per day (cap)
Hours
Per candidate (read, match, write)
100%
Recruiter time on screening + outreach
The consequence
No time for the actual recruiter work: holding conversations, building relationships, guiding candidates. The day went to searching, not meeting.
02, Process flow
Meet Lara
Lara is the AI coworker who owns the entire sourcing process, from pulling profiles to dropping the InMail in the candidate's inbox. Recruiters get capacity and focus back.
01
Retrieve Full Profile
Lara pulls the full candidate profile, work history, skills, public projects. Not just what's on a CV, but the complete picture.
02
Match to vacancies
The profile is matched against open roles: hard requirements, soft requirements, contextual fit. Only true matches move on to the next step.
03
Compose InMail
For every strong match, Lara writes a personal message, no template, but an invitation that references what the candidate does and why the role fits.
04
Send InMail to Mailbox
The message is sent. The recruiter has real-time visibility into the pipeline and can focus on responses instead of searching.
05
Human-in-the-loop on uncertainty
On 2.31% of cases a recruiter steps in, for ambiguous matches or profiles that need context. The platform learns from every intervention.
24,298
Candidates in the pipeline, all screened
~4,500
Match against open vacancies
1,126
Personal InMails sent
The funnel, Q1 2026
03, Results
Live production data from Q1 2026
Measured across 24,298 processed cases between January 1 and March 31, 2026.
KPI
Realized
Context
Delta
Candidates screened
24,298
Q1 2026
Personal InMails sent
1,126
Top matches
Handling time per candidate
7 sec
End-to-end
Throughput time
1d 1h 34m
Includes external wait
Human involvement
2.31%
On ambiguous matches
Exception rate
0.43%
Stable, low
Note on volume.InMails sent (1,126) is deliberately lower than the number of screenings (24,298). Lara optimizes for precision, not quantity, only true matches get a personal outreach. The 22,000+ uninvited candidates aren't rejected; they just aren't the best fit for the current openings.
04, What this really means
Recruiters who get to be recruiters.
BPM Company's story illustrates something every recruitment team faces: the heavy lifting isn't the conversation with a candidate, it's finding that candidate. Combing through lists, comparing profiles, writing personal messages, hours of work that doesn't scale.
With Lara the work shifts: the AI does the screening, recruiters do the conversations. No more 30-a-day cap; no more good matches lost in the pile because there simply wasn't time. The pipeline stays continuously full, and every invitation that goes out is deliberate.
For recruitment teams the effect is twofold: more quality per outreach, and more time for the work only humans can do, listening, persuading, guiding.
Own recruitment flow?
Find out what an AI coworker can do for your team.
We start with a Quick Scan. No commitments, just a concrete plan on the table within two weeks.