AI Recruitment

Hire the best — in half the time

An AI-powered ATS, from job posting to offer letter.

Nova HR reads your résumés, ranks candidates against the job description, and turns weeks of manual screening into minutes of informed decisions. Hire with confidence, not guesswork.

Role: Full-Stack Developer

148 applicants — AI-analyzed

Live
  • SA

    Sara Ahmed

    Full-Stack Developer

    ReactNode.jsPostgreSQL
    96%
  • MH

    Mahmoud Hassan

    Full-Stack Developer

    ReactPythonAWS
    88%
  • NK

    Noor Kamal

    Full-Stack Developer

    VueDjangoDocker
    74%
Candidates auto-ranked by match score11 hours saved
AI-powered ATS

Hire the best candidates — in half the time

Nova HR ships with a built-in AI-powered ATS that reads résumés, ranks candidates by fit, and turns hiring from weeks of manual screening into minutes of informed decisions.

  • Smart CV parserPulls data out of any PDF or Word résumé in seconds.
  • Smart candidate matchingRanks candidates by skills, experience, and job-description fit.
  • End-to-end hiring automationFrom posting to offer letter — automated stages, notifications, and approvals.
Everything you need

A full hiring pipeline — powered by AI

From posting a role, to screening applicants, to interviews, to the final offer.

AI CV parser

Extracts experience, skills, and education from any PDF or Word in seconds.

Smart candidate matching

Scores candidates against the job description and ranks them automatically.

Multi-channel job posting

Post jobs once to LinkedIn, Bayt.com, and more at the same time.

Boolean search & filters

Boolean-search your candidate pool and build shortlists in minutes.

Interview scheduling w/ Zoom

Schedule one-on-one and panel video interviews with Zoom built-in.

Branded careers page

A careers page under your domain, fully branded — no code required.

How it works

How recruitment works in Nova HR

From the hiring request to the employee file — the funnel in detail, and what the AI actually does in it.

The hiring request: the need before the ad

Most hiring starts with a message: "we need an accountant". That beginning already contains everything that will go wrong two months later, when a candidate is ready and asks for a figure nobody approved.

So the cycle starts with a request recorded by the department with the need, passing review and approval before it becomes a published ad. It carries the position, the department, who the new hire reports to, the required start date, the number of vacancies, how many people will report to the role, whether it is a new role or a replacement, and the budgeted salary range.

The third is the most revealing: a replacement is a cost already budgeted, while a new role is an increase in headcount and annual cost. Companies that do not distinguish them discover their growth at year end rather than during the year — because each hire looked like a reasonable replacement at the time.

AI screening: what the AI actually does

One ad brings hundreds of CVs and reading them all carefully is impossible. What happens is that the first twenty are read with attention and the rest are scanned for a keyword — and good candidates are lost because their CV arrived late, not because it was weaker.

The precise description of what the AI does is not "it picks candidates" but two things: it reads and it ranks. As soon as the CV is uploaded — a document or an image — it extracts the data and auto-fills the application form, removing the manual copying step entirely.

Data extraction
name, email, phone, qualification, skills, experience and courses — filled straight from the CV.
Match percentage
a figure from zero to a hundred for how well the CV matches the advertised job's requirements.
A match summary
a short paragraph in Arabic and English summarising strengths and gaps against the role.
A per-requirement verdict
met or not met, with a brief reason drawn from the CV itself — which is what makes the score judgeable.

Why the AI does not decide

A CV is an incomplete document by nature. It does not carry the real reason for leaving a job, the quality of work behind a title, or the circumstance explaining a gap. Judging those is precisely what distinguishes one candidate from another with identical credentials.

Leaving the decision to a match score produces a systematic bias toward whoever writes their CV in the ad's own language — a writing skill, not a work skill. So after the automated output the CV passes three sequential human review stages, each performed by a different reviewer.

Whoever does not fit the current role is not discarded: their CV is kept in the CV bank for future openings. What the company paid to reach them stays an asset that compounds, and the tenth role starts with known candidates instead of starting from zero like the first.

Interviews: from call status to assessment

Between screening and the interview lies the space where candidates get lost: called and no answer, promised to come and did not, asked for another date. Without tracking these states the recruiting team asks itself every morning who is left and who dropped out.

So the call status is recorded with the interview time and call notes, then two separate results: the HR interview assessing behaviour and general fit, and the technical interview assessing ability to do the job. The separation is deliberate — a candidate who passes the technical and fails the behavioural is an entirely different decision from the reverse.

The offer: four decisions, not one number

The common reduction is that an offer is the salary. It is in fact four decisions: the contract type that sets its term and what follows from ending it, the salary structure, the package breakdown into basic, variable and commission, and the start date.

The package breakdown matters more than it looks: the difference between a gross split into basic and allowances and the same gross split differently is a real difference in insurance, tax and end-of-service gratuity — not a cosmetic one. A candidate signing for a "gross" without knowing its split will ask about the difference on their first payslip, and by then the question is an objection rather than an enquiry.

The offer prints on an official letter template, and candidates' replies are tracked from the offer-statuses screen — because every day in that void is a day the role goes unfilled, and a day the candidate may accept another offer.

Hiring: the funnel's end is the employee file's start

The moment a good hiring process most wastes its own quality is the moment it succeeds: the candidate accepts, and HR starts entering their data from scratch as though the previous two months never happened.

The hiring screen shows only accepted offers whose holders are not yet employed, and the start date and identity data are entered. Identity is mandatory except in one case: hiring into a training programme, where the person is recorded as a trainee and it becomes optional.

And in one step all of this happens together: the employee file opens with an auto code, their branch, job and reporting line; marital status, religion and military status carry over from the applicant's form; contact details, address and experience are copied; a login is created with default permissions; and the offer is marked as hired so it drops off the list.

Outcomes

What you actually get

Not just features — measurable outcomes you'll see in your company within weeks.

  • Cut time-to-hire by 50% — from 4 weeks to 2.
  • Zero manual résumé review — AI summarizes and scores fit.
  • Make data-driven hiring decisions, not gut calls.
  • A searchable candidate pool that compounds with every posting.
  • A professional candidate experience that lifts offer-acceptance rates.

Common questions

Yes, and this is one case where an automated reader beats keyword filtering: the match rests on understanding the content rather than textual identity. It is a practical matter in a market where CVs arrive in both languages, sometimes mixed within the same document.

No single number suits every role, and it is better not used as a hard cut-off at all. Using it to order is more useful: read from the top down until you have enough candidates for interviews. A hard threshold automatically excludes people who may deserve an interview, and the gap between two close scores is not a real gap.

The usual objection is that approval slows hiring. What actually slows it is rework: an ad pulled because the budget was never allocated, and an offer delayed two weeks awaiting a sign-off that should have been taken two months earlier. Prior approval moves the waiting to the start of the path, where no candidate is waiting.

Their CVs are kept in the searchable CV bank and can be revisited whenever a role opens. The most wasted thing in recruiting is not time but the CVs collected and then abandoned — someone who did not fit one role may fit another months later, and reaching them cost a paid ad.

Yes, if each applicant's source is recorded accurately. That is what later reveals which channel brings the best candidates — not the most — so budget goes there instead of being split evenly across channels whose results are not even.

No. What the candidate wrote about themselves on the application travels with them through the whole path and lands in their file: personal details, contact, address and experience. A login is created automatically with default permissions — so accuracy at the start saves re-entry at the end.

Start with Nova HR today — one step.

Book a demo tailored to your company — 30 minutes with an HR consultant showing exactly how every module fits into your workflow. No credit card. No commitment.

Or reach out at info@dynamiceg.com