Three Days to Same-Day: Automating a Hiring Pipeline
May 22, 2026
- Problem
- A candidate filled in a form, the submission arrived as an email, and someone re-typed it into the applicant tracking system. Screening then happened when somebody had time — about three days, at 200+ applicants a month.
- Constraint
- Manatal's default pipeline described a hiring process nobody at the company actually followed, and reference checks were a sequence of small, easily-forgotten follow-ups rather than a hard technical problem.
- Decision
- Mapped the real process with the people doing the hiring before configuring anything, then had the WordPress careers form post straight into Manatal over its REST API instead of into an inbox. n8n runs the background and reference checks on a schedule rather than on someone's memory.
- Outcome
- Screening fell from three days to same-day, manual data entry dropped by 80%, and reference-check response rates rose 25% across 100+ candidates a week.
When I picked this up, applying for a job here worked like this: a candidate filled in a form, the submission arrived as an email, and someone on the team re-typed the details into the applicant tracking system. Then screening happened when somebody had time.
End to end, a candidate waited about three days to hear whether they'd move forward. At 200+ applicants a month, the re-typing alone was close to a part-time job.
Three things changed that.
1. Making the ATS match the actual process
We used Manatal. Out of the box it has a sensible default pipeline, and the default was wrong for us — it described a hiring process nobody here followed.
So the first work wasn't technical at all. I sat with the people doing the hiring and mapped what they actually did: which stages existed, who decided what at each one, what caused a candidate to stall. Then I configured Manatal to look like that.
This is the step that's easy to skip and expensive to skip. Automating a process nobody follows produces a system everybody works around.
2. Deleting the re-typing step
The application form lives on the careers site, built in WordPress with Elementor. Rather than have it send an email that a human then transcribes, the form posts directly into Manatal through its REST API.
A candidate submits, and the record exists in the ATS. No inbox, no copy-paste, no transcription typo in someone's surname.
That one change removed about 80% of the manual data entry in the whole process. It isn't a clever piece of engineering — it's an API call. The value was in noticing that a human was being used as an integration layer.
3. Chasing references without chasing references
Background and reference checks were the remaining bottleneck — not because they're hard, but because they're a sequence of small, easily-forgotten follow-ups.
I built the sequence in n8n: when a candidate reaches the relevant stage, requests go out automatically, responses are tracked, and non-responses get chased on a schedule rather than when someone remembers.
It now handles 100+ candidates a week, and response rates went up 25% — almost entirely because the follow-up reliably happens now.
Where it landed
- Screening: 3 days → same-day
- Manual data entry: down 80%
- Reference-check response rate: up 25%
- 100+ candidates processed weekly without manual chasing
The part I'd emphasise
None of this was difficult engineering. There's no clever algorithm — it's platform configuration, one REST integration, and a workflow tool.
The work that mattered was mapping the real process before automating it, and spotting that a person was doing a computer's job in the middle of it. Both of those are conversations, not code. I think that's true of most automation worth doing.