The introduction of AI into recruitment has had a massive—and likely irreversible—impact on both recruiters and job seekers.
Qarar Research Team
The Qarar Research Team gathers and analyzes current research and data in HR and recruitment, to deliver content grounded in real evidence, not opinion or impression.
The introduction of AI into recruitment has had a massive—and likely irreversible—impact on both recruiters and job seekers. At the same time, the needs of companies themselves are changing radically as AI becomes increasingly integrated into their internal operations.
In this piece, we demonstrate:
Why AI is changing the recruitment market and why the old metrics are becoming less reliable.
Where AI actually belongs in the hiring process.
How recruiters can use AI without surrendering human judgement.
How to design better job descriptions and application forms.
How to screen candidates without falling into the keyword trap.
How to automate sourcing, communication, verification, and talent-pool management.
How to build a recruitment process that is faster, more data-driven, and still meritocratic.
How to win the AI recruitment war.
Recruiters can now use AI to generate job advertisements, publish vacancies across multiple job boards simultaneously, and source candidates through agents that crawl websites and extract candidate information, often with limited regard for privacy and individual autonomy. AI is also increasingly being used to scan, score, and rank candidates according to predefined criteria.
But these systems are not necessarily objective.
Large language models can interpret the same information differently depending on the model, prompt, context, and system being used. This introduces a degree of stochasticity into processes that are often presented as if they were precise and deterministic.
You can share the article
An optimisation war has consequently emerged.
Applicants are learning to produce increasingly inflated CVs packed with keywords designed to satisfy automated screening systems. Recruiters, in response, are becoming less dependent on simple keyword matching and are developing more sophisticated contextual and dynamic scoring mechanisms.
The more sophisticated one side becomes, the more sophisticated the other must become.
At the same time, the emergence of AI-generated applications and other forms of application fraud threatens to undermine one of recruitment’s most basic assumptions: that an application represents a genuine effort by a genuine candidate. Research increasingly points towards a future in which a significant proportion of applications may be artificially generated or otherwise inauthentic.
If that happens, application volume and many of the other superficial metrics traditionally used to evaluate recruitment performance, will become increasingly meaningless.
Where AI belongs
We believe AI has an important place in the hiring process. But it should not be an alternative to human judgement, nor should it become a mechanism that rewards manipulation and undermines meritocracy.
AI should be used primarily for what it does best: repetitive, logistical, administrative, and information-processing work.
It can fill forms, send emails, organise candidate information, filter applications, coordinate communication, and generate statistics that help decision-makers reach better, more data-informed conclusions.
The machine should reduce the administrative burden.
The human should remain responsible for the decision.
How to Win the War
1. Protect your eyes and your brain.
Your own judgement is still one of the most trustworthy tools you have. Do not completely delegate it to an unreliable machine. AI can help you process information. It should not replace your ability to understand people.
2. Write the job description yourself.
Start with a clear understanding of the role, its responsibilities, and what success actually looks like. AI can help you phrase and refine the description, but the underlying definition of the role should come from the people who actually understand the work.
3. Do not rely heavily on keywords.
Keywords can be artificially inflated. A candidate mentioning “leadership,” “Python,” “strategic planning,” or “project management” repeatedly does not necessarily possess those capabilities. Look for evidence, context, and relevance rather than vocabulary.
4. Use the right tools for sourcing, screening, and scoring.
AI can dramatically reduce the amount of manual work involved in recruitment. Use it to search, organise, filter, and prioritise candidates—but understand what the system is actually measuring before allowing its output to influence your decisions.
5. Do not ask candidates questions you already have the answers to.
If the CV already contains the information, there is little value in forcing the candidate to enter it again into an application form. Every unnecessary question adds friction and creates another opportunity for candidates to abandon the process.
6. Ask the questions that actually matter.
Think carefully about the few questions that can meaningfully distinguish candidates at the screening stage. Put those questions in the application form. A well-designed application form should allow a recruiter to understand the essential characteristics of a candidate in seconds, rather than requiring them to reconstruct the candidate’s profile from dozens of irrelevant fields.
7. Never leave candidates without an answer.
Use an ATS with automated communication so candidates receive an outcome when their application is no longer being considered. A rejection email may be disappointing, but it is more respectful than silence. Good communication builds trust in the employer and protects the reputation of the recruitment process.
8. Build a progressive talent pool.
Some ATS platforms allow organisations to maintain their own talent pools. When a strong candidate is not selected for one position, that does not mean they are irrelevant forever. Tag and retain promising candidates so that they can be contacted when a relevant position becomes available.
9. Verify identity and credentials early.
Identity and credential verification should happen near the beginning of the hiring process—not after significant amounts of time and resources have already been invested. There is little value in optimising a candidate through an entire recruitment funnel only to discover at the end that their identity, qualifications, or experience cannot be verified.
10. Record remarkable professional behaviour.
Recruitment should not be reduced to CVs and test scores. Remarkably positive—or remarkably negative—professional behaviour can be valuable information for future hiring decisions, provided that such records are factual, relevant, consistently applied, and handled in accordance with applicable privacy and employment laws. A candidate who demonstrated exceptional professionalism during a previous recruitment process may be worth remembering. So may a candidate who repeatedly demonstrated behaviour that raises legitimate professional concerns.
The objective
The objective is not to build a recruitment process where AI makes all the decisions.
It is to build one where AI removes everything that prevents humans from making better decisions.
The winning recruiter will not necessarily be the one with the most sophisticated AI.
It will be the recruiter who understands what to delegate, what to automate, what to measure, and what should never be delegated at all.
