HR Automation in Arabic: Screening, Policy and Nitaqat

HR automation in Saudi Arabia and the UAE works when the system does the reading, chasing and tracking, and a person makes every hiring decision. Built Arabic-first on your handbook and your role scorecards, it returns hundreds of hours a year without moving a single decision out of HR's hands.

Most Gulf HR teams do not have a technology problem. They have a memory problem. The policy answer lives in the HR lead's head, the expiry dates live in a spreadsheet one person updates, and the nationalisation ratio gets checked after the offer is signed instead of before.

Three jobs to hand over first

Start with the work that is repeatable, measurable and nobody's favourite part of the job. Three jobs qualify in almost every HR team we see.

1. CV screening against a written scorecard

Every CV is read against the role scorecard, and the reason for each shortlist decision is written down. HR reviews the shortlist with the reasons in front of them, instead of reading every CV from the top.

The arithmetic, on stated assumptions: 250 CVs per role, 4 minutes each, 20 roles a year. That is about 330 hours of reading a year, returned to the people who should be interviewing.

2. Policy answers from the handbook, in two languages

The same question about leave, allowances or probation arrives forty times a quarter, in Arabic and in English. The system answers from your handbook, quotes the clause, and routes anything the handbook does not cover to HR.

3. Expiry and ratio tracking

Visa, work permit and contract expiries are tracked 90 days ahead, with the reminder going to the right owner. For Saudi employers, the Nitaqat band the Ministry of Human Resources and Social Development assigns is tracked before a hire changes it; in the UAE, the same applies to the Emiratisation targets set by MoHRE.

Why Arabic-first is a requirement, not a feature

Half the inbox is in Arabic. Employees ask policy questions in the language they think in, and CVs arrive in both. A system that reads English well and Arabic as an afterthought misreads half the work and answers the other half in the wrong register.

Arabic-first means the handbook, the scorecards and the answers are built in both languages to the same standard, and the answer comes back in the language the question was asked in.

What stays human

• No autonomous hiring or firing decision. The system shortlists with reasons. A person decides.
• Screening criteria are audited for bias before they run, and every reason is visible to the person who approves the shortlist.
• Offers come from approved templates and salary bands, and a person releases them.
• Employee data stays in your systems, under your data-protection obligations in KSA and the UAE.

How it rolls out

The same three stages as any AI-native operating system, one job at a time. Applied: the HR team uses the scorecard and the handbook assistant on its own work. Agentic: the system screens and drafts, HR approves every shortlist and every answer it has not seen before. Autonomous: expiry reminders and standard policy answers run on their own, and only exceptions reach a person.

Each job is one process: one trigger, one output, one owner. Built, run on real work and handed to your HR lead in two weeks, at $7,500 per process.