One hiring system, three times the candidates screened
A custom recruiting system that puts every applicant in one searchable place, with stages, notes and handoffs, and shows the company where hiring stalls.
Recruiting ran on a Google Form and one job board. Applications were exported to a spreadsheet and sifted by hand by two or three recruiters. There was no standard way to view an applicant, notes got lost, and there was no way to pause and hand a candidate to a colleague. If someone got busy, the candidate simply waited, and waiting candidates get lost. About three or four candidates were screened a week.
Hiring is the engine of a BPO: they hire to meet their clients' needs. If they cannot hire fast enough, they lose clients, revenue and growth.
The COO asked for it, with HR's input. The form and the job board were not the problem. The problem was sifting through everything, so that is what we aimed at. HR and recruiters needed:
- One standard place to see every applicant: the resume, the answers, and the details pulled out of them, all searchable and filterable.
- The video, stages, notes and who did what, in the same view, so anyone could pick up where someone left off.
- A clear path from applied to screened, through interviews, to offer and hired, even when one posting means several hires.
- A view of how hiring was going, now and over its whole history.
Two ground rules shaped the design. Candidate details are private: only HR sees them, and everyone else builds on test and anonymized data. And a human always decides, because software should help people do what they do better, not replace their judgment.
The budget was a requirement too. As a growing company, they needed their money in their most important resource, their people. So the system had to fit their cost needs, today and in the near future.
We did not buy an expensive suite or replace everything. We built one bespoke, extensible system aimed only at the part that was broken: sifting through applicants.
The Google Form and the job board. They already worked, so applications still arrive the same way.
The spreadsheet and the manual sifting. A scheduled job reads every CV, extracts the details, and stores them in one database.
One screen for recruiters: resume, extracted details, video, stages and shared notes. Every action is logged and audited.
Durable and reliable, with alerts and monitoring, autoscaling, and tuning that balances processing time against cost.
Custom fit to their needs today and in the near future, and built to be extended as new requirements come.
Proper, detailed documentation left with the company, so they can run it and keep growing it without me.
Processing runs in batches every few hours, not instantly. That keeps the running cost down, and recruiters were never reacting in real time, so they never noticed.
I led nine AI-equipped developers in three teams, each with a clear job and clear contracts between them:
I set the foundation and the rules first, so nine developers could move fast without getting in each other's way.
Recruiters screened about three times as many candidates and were very happy with the system. Notes stopped getting lost, candidates stopped waiting on one busy person, and the company could finally see its hiring funnel. The analytics showed that in the summer surge, both years, they screened fast enough but interviewed too slowly. The bottleneck had moved from sifting to interviews, which is a much better problem to have.
Why it worked: we fixed the part that was actually broken, built it to fit and to grow, and kept a human in charge of every decision.
Hiring stuck in forms and spreadsheets?
Book a free 30-minute call and we will see whether a lean, custom fix makes sense.