Who this is for
Manpower and overseas recruitment agencies processing high application volumes for a small number of positions, and in-house HR teams hiring steadily enough that the administration has become the job. It is one of the builds under marketing and business automation services. The related industry page is automation for manpower and recruitment.
The volumes are what make it worth doing. When four hundred people apply for twelve positions, no amount of diligence lets a two-person team read every CV properly, respond to everyone and still run the interviews. Automation buys back the hours so the judgement happens where it matters.
What gets automated
- One application intake. A single form behind every channel, so applications from Facebook, the website, a job board or a WhatsApp number all arrive in the same structured record with the same fields, instead of as forty attachments in an inbox.
- CV parsing. Name, contact, experience, skills, languages, certifications and availability extracted into columns so a shortlist can actually be sorted, with anything the extraction is unsure about flagged rather than guessed.
- Rule-based screening. Hard requirements you define, such as a valid passport, a specific trade certificate, a minimum experience or a language level, applied consistently. Everyone who fails a hard requirement gets a prompt, polite answer instead of silence.
- Acknowledgement and status updates. Every applicant knows their application arrived, roughly when they will hear, and what happens next. This is the cheapest reputational improvement most agencies can make.
- Interview scheduling. Shortlisted candidates pick a slot from the interviewer’s real availability, with reminders and easy rescheduling, using the same mechanism as appointment booking and reminder automation.
- Document collection. The checklist per candidate, chased automatically, with each document checked for the obvious problems such as an expired date or a wrong file type before a person opens it.
- Onboarding. Offer letters generated from a template, accounts requested, first-week checklists issued and the new joiner kept informed between signing and starting.
- Reporting. Time to hire, source of successful candidates, drop-off stage and cost per hire, so recruitment advertising can be judged on hires rather than applications.
What I deliberately do not automate
- Hiring decisions. Ranking, scoring and rejecting on judgement stays with people. Automated screening applies only the hard requirements you wrote down, and a person can see and overturn every filtered application.
- Personality and video assessments scored by a model. The evidence for these is weak and the discrimination risk is real.
- Anything that filters on age, gender, caste, ethnicity, religion, marital status or region, directly or by proxy through a field that stands in for them.
- Charging candidates. I do not build payment collection from job seekers into recruitment workflows.
- Storing passports and identity documents in spreadsheets. These stay in access-controlled storage with a retention period, and the workflow handles references to them rather than copies. Have your own adviser check the consent wording and retention period against Nepal’s labour and foreign-employment rules before the workflow goes live.
How a build runs
- Write the requirements down properly. Which criteria are genuinely hard requirements and which are preferences. Most teams have never separated the two, and doing it is half the value of the project.
- Design the candidate journey from application to rejection or offer, including the message at each stage and the maximum time a candidate should wait for it.
- Build the intake and parsing and test it on two hundred real past CVs, measuring how often extraction and screening agree with what your team decided. That number decides how much of the screening runs unattended.
- Set the data rules. Who can see what, where documents are stored, how long records are kept and how a candidate’s data is removed on request.
- Pilot on one open position, with a person reviewing every automated decision, then widen once the review is consistently agreeing.
- Hand over with documentation, training for the hiring team and a monthly report on time to hire and drop-off.
What you receive
- The application pipeline in your own accounts, with candidate data in storage you control.
- A written screening specification: every criterion, whether it is hard or preferred, and who approved it.
- Candidate message templates for every stage, including a rejection that is respectful and specific enough to be useful.
- Role-based access, a documented retention period and a process for deletion requests.
- Hiring reports, error alerts, documentation and a training session for the recruitment team.
Pricing pointer
Setup is quoted on request; an intake, parsing and status-update pipeline for one hiring process sits at the lower end of the range. A full agency system covering multiple job orders, document collection, interview scheduling and reporting sits at the top. Maintenance is quoted on request; applicant tracking or storage subscriptions are paid in your own name. Where recruitment advertising is the bottleneck rather than the processing, see Google Ads for manpower and recruitment.
Next step
Tell me how many applications a typical position attracts, where they arrive from and what your hard requirements are. I will come back within four business hours with the pipeline I would build, what would stay manual and a fixed price.
