Today’s question is no longer whether we should use AI in recruitment. It has already become a reality in HR, globally and locally. For example, 99% of Fortune 500 companies rely on AI-powered Applicant Tracking Systems (ATS), where résumés are automatically screened before they ever reach recruiters.

But the clear truth is this: it’s difficult to unlock the full value of AI capabilities unless we know how to direct it through clear, structured instructions.

These instructions should also align with the nature of each hiring stage, while remaining compliant with labor laws across the Middle East. So the real question becomes: How do you use AI to truly save hours of manual work?

But first, let’s answer this: Why do AI results fail?

The Top 3 Reasons AI Outputs Fail (Prompt Failures)

If your requests have any of the traits below, AI will likely produce inaccurate or unsuitable outputs:

1. Ambiguous language: It is expected to get results that don’t fit your needs when you use terms that are vague or have multiple meanings.
2. Unclear context: AI struggles to tailor its response to your actual need when it lacks sufficient background, such as the goal of your request or the nature of the role.
3. Overly complex requests: Requests that are too complex, or combine multiple tasks in a single instruction, often lead to scattered or incomplete outputs.

So, how do you give AI a clear instruction across different recruitment workflows?

To answer that, let’s start with…

The Core Elements of a High-Precision Prompt

AI users typically rely on several elements to improve accuracy and output quality, including, but not limited to, the following:

• Task: Define exactly what you want from AI. The clearer and more direct the task, the more precise and useful the output.
• Persona: Specify the role AI should “assume” while completing the task. Think of it as casting a role for an actor within a clear context.
• Context: Provide background that helps AI understand your request accurately, such as the objective, target candidates, hiring stage, and geographic location.
• Format: Clarify the desired output format. Do you want bullet points, a detailed description, or a structured table? Specifying format reduces distraction and prevents exhausting manual edits.
• Example / Model: Providing a reference example can be one of the strongest guidance tools, encouraging AI to match your style, depth, and expected output quality.
• Tone: Set the appropriate tone for recruitment content, formal, marketing-style, or even a light tone, depending on the situation.

Practical Examples

Consider the following prompts:

Example 1
“You are the Digital Marketing Team Manager at a Gulf-based company [insert persona name or description]. Write a concise email [task] addressed to the executive leadership, summarizing the results of the latest recruitment campaign [context], based on this report [attach a clear example]. Use a formal professional tone [tone], and present the points as a brief bullet list [format].”

Example 2
“I will provide you with an example of an interview question for a Saudi digital marketing specialist [context]. I want you to generate 10 similar questions [task]. Here is the example [model]: ‘Can you describe a situation where you had to change a marketing plan under tight time pressure?’”

Notes

Do you need to use all of these elements at once?

The short answer: No. You can use only what fits your request, or use some elements in the first prompt, then follow up with additional instructions that include the rest. But remember: the most important element in any prompt is the task. You must be precise and practical when defining what you want AI to do.

Also remember: you may add other elements beyond those listed above, such as the desired outcome of the task. In the last example, you could add:
“Create a set of high-quality interview questions aligned with the cultural context and the Saudi market, designed to uncover the candidate’s thinking and adaptability.”

Now that we’ve covered the basics, let’s apply these principles across different hiring stages.

Using AI in HR

1) Writing Job Descriptions

If you’re using AI to create a job description, you must clearly specify the job title, key responsibilities, required qualifications, and any additional advantages you want to highlight.

It is recommended to define the tone in line with your employer brand, and to instruct AI to write in your preferred format. If you have an ideal job description model, share it and ask the system to follow the same style.

Practical Example

A sample prompt for a Digital Marketing Specialist job description:

“You are an HR expert working at a leading Saudi technology company. Write a comprehensive job description for the role of ‘Digital Marketing Specialist,’ and explain how it supports brand growth and demand generation for digital products. Focus on responsibilities such as managing digital campaigns, analyzing performance, and collaborating with content and sales teams. Include required qualifications such as a Bachelor’s degree in Marketing or a related field, at least 3 years of experience in digital marketing, and proficiency with analytics tools and advertising platforms. Also highlight the company culture based on innovation and market impact in Saudi Arabia.”

If AI fails to produce the job description in the style you prefer, guide it with a reference model, such as:
“Analyze this job description, then rewrite a Digital Marketing Specialist job description with the same depth and style, adapted to the Saudi market. The model to emulate is: (attach the example here).”

Additionally, explicitly ask AI to balance legal compliance for the local labor market with language that appeals to experienced candidates, while avoiding vague phrases that don’t reflect the real role.

However, you can avoid much of this effort by using SANAD, the first smart hiring assistant in the Middle East, embedded inside Talentera, which provides, among many other things, a professional job description in seconds.

2) Creating Job Ads

When moving from a job description to a job ad using AI, focus on persuasion and selection at the same time. A strong job ad doesn’t only explain the role, it also communicates your company culture and highlights it to the right candidates. It helps to instruct AI based on the target candidate’s behavior and the channel where the ad will be published.

It’s also useful to mentally structure the ad into three layers:

1. An opening line that captures attention in the first seconds
2. A value paragraph explaining why the role matters and how it fits your organization
3. Key details that focus on expected impact, not a long list of tasks

You may also ask AI to generate multiple versions of the same ad depending on the channel, such as a short version for fast platforms and a more detailed version for serious candidates or specialized job boards. (And don’t forget to instruct it to comply with local regulations and avoid overused clichés that weaken credibility.)

A sharp recommendation: One of the best ways to improve AI output quality is to ask it to evaluate the job ad after writing it, identify possible turn-offs and suggest edits that attract top talent.

Practical Example

A sample prompt for a job ad:

“You are a Talent Acquisition specialist [persona] working in a mid-sized Saudi tech company going through a focused growth phase [context]. Your task is to write a high-quality professional job ad for a ‘Digital Marketing Specialist’ role [specific role], targeting professionals with real hands-on experience in the Saudi market who are looking for a meaningful challenge and real ownership [target audience]. Write in a professional tone, avoiding generic recruiting language and overused phrases, speak as if you’re addressing a candidate who understands digital marketing deeply and doesn’t need the basics explained [tone]. Start with a paragraph explaining why this role matters specifically in our company, and clarify the value expected from the candidate in the first 3 to 6 months in paid campaigns, performance analysis, and collaboration with content and sales teams [specific requirements].”

Note

You may want to split the above instructions into stages depending on AI’s response quality. You might also add details such as requirements, assessments, and cultural points, based on the target audience, while maintaining compliance and credibility.

3) Writing Candidate Emails

Email remains one of the most-used channels for candidate communication, whether to share information or send updates across hiring stages. The real challenge isn’t sending emails, but managing them efficiently, especially when HR teams handle dozens, or hundreds, daily.

To reduce this load and speed up candidate communication, include these elements in your email prompts:

• Reason for the email: Don’t only specify what to write, explain why. (To find the reason, ask: Is the goal to reassure the candidate, encourage action, or reduce follow-up questions?)
• Hiring stage: Clarify where the candidate currently is in the process.
• Tone and emotional impact: Instead of saying “write professionally,” say “write so the candidate feels reassured and appreciated, while reflecting seriousness and personal attention.”

Practical Examples

Early outreach email:
“Draft an initial outreach email to a potential candidate for the role of [job title]. Highlight why we selected them specifically based on strengths in [relevant skills], then briefly explain the role’s value and professional challenges, and clearly outline the next steps [provide a roadmap].”

Interview invitation email:
“Write an interview invitation email for the role of PR Manager [job title]. Include the date/time, interview format, attendees, and any preparations [provide details]. Then explain what the candidate can expect, question type and meeting duration. Use a professional, reassuring tone.”

Final evaluation stage follow-up:
“Create a follow-up email template after interviews for candidates in the final evaluation stage. Express appreciation for their time and effort, clearly state the next steps and expected timeline, and use a human tone that reassures candidates and reflects respect regardless of the final decision.”

Rejection email:
“Write a rejection email to a candidate who interviewed for [job title] and did not pass the screening stage. The message must be polite, reassuring, and respectful of the effort they invested, leaving a positive impression that preserves a future relationship. Optionally include general feedback aligned with [insert your preferred recommendations, if any].”

The heart of it: The purpose of AI in candidate messaging is to reduce bias, standardize communication, and speed up updates. But this process is not complete without a professional human review that preserves your company voice. Read templates carefully and adjust key phrases as needed.

5) Résumé Screening

Some ATS platforms, such as Talentera, enable AI-powered résumé screening and ranking. The system analyzes profiles based on qualifications, core experience, technical and behavioral skills, and other relevant factors, then ranks the most suitable candidates against your role requirements.

To maximize this capability, take the following steps before writing AI prompts:

• Know your role requirements and priorities: If you cannot define your priorities, every résumé will distract you. Clearly define years of experience and required credentials, and share them with AI (while considering local professional regulations where relevant).
• Define what AI may miss: A high rank doesn’t guarantee soft skills or cultural alignment. Strengthen AI recommendations with later-stage filters such as assessments.

Practical Example

“You are a Talent Acquisition specialist at a Saudi tech company. You have a set of résumés for candidates applying to a Digital Marketing Specialist role in the Saudi market. Extract core experiences, education, technical skills, and behavioral skills for each candidate.”

You can add these instructions:

• “Evaluate each candidate’s fit with the role, focusing on campaign management and performance analysis.”
• “Identify each candidate’s strengths, especially Saudi market experience.”
• “Highlight weaknesses or potential gaps without absolute judgments.”
• “Rank candidates by years of experience and alignment with role requirements, and provide a professional justification for the top 10 selections.”

A focused recommendation: Break your prompts into sequential stages to improve output quality.

Note

AI can also help you first define the best screening criteria, not only screen candidates. For example:

“You are a Talent Acquisition specialist in a mid-sized Saudi organization. Define a clear, detailed set of screening criteria for evaluating résumés for the role of [job title]. Specify the qualifications, skills, and experiences that should be prioritized, and include preferred (nice-to-have) criteria.”

6) Conducting Market Research

AI can make market research easier by quickly surfacing resources and summarizing recruitment trends, helping you understand demand, salary levels, competitor behavior, hiring approaches, reports, and real skill gaps.

But the quality of these outputs depends directly on source quality and prompt precision. Always define the timeframe, tie it to a geographic location, and review sources to verify reliability.

Practical Recommendations

• Define the research scope: Don’t ask broadly for “recruitment market research.” Specify the job sector, country or city, target company size, and role characteristics.
• Clarify what decisions the research will support: Ask yourself:
  • Is the goal salary benchmarking?
  • Measuring sourcing difficulty?
  • Redesigning the job description?

Note: Be aware of AI limitations, some qualitative signals (employer reputation, cultural candidate preferences) may require human interpretation.

Practical Example

“You are an HR Manager at a large Saudi financial company. Conduct market research on hiring a Financial Systems Analyst within financial services in Saudi Arabia for 2026, considering regulatory and compliance requirements. Focus on current salary trends for a mid-level Financial Systems Analyst, and highlight current high-demand skills as well as skills expected to grow in demand over the next two years.”

Notes That Matter

• Use AI to generate the key questions that reveal market realities, don’t limit it to answering only your pre-written questions.
• Ask AI to provide the direct sources it relied on, then review them for inconsistencies or inaccuracies.
• Request structured analysis or comparison tables depending on your needs, and, as always, splitting prompts into stages improves quality.

7) Generating Interview Questions

High-quality interview questions bring you closer to candidates and clarify motivations, strengths, and challenges. AI can help generate dozens of role-specific questions, but the output mirrors the maturity of your prompt. The more clearly you define the role context, real challenges, and your company culture, the more realistic and accurate the questions will be.

Fast Recommendations

• Specify interview type and define interviewer roles in panel interviews.
• Don’t only mention the title, define the candidate’s experience level.
• Explain your company stage (growth, restructuring, digital transformation, etc.).
• Mention common hiring mistakes you want to avoid, and define expected autonomy and impact in year one.
• Ask AI to explain the purpose of each proposed question to ensure relevance.

Practical Example

“You are an HR Manager at a Saudi financial company undergoing digital transformation. Create interview questions for a Financial Systems Analyst role. Design questions to cover the ability to analyze financial requirements and translate them into practical solutions (focus on the core capabilities to uncover), as well as questions that reveal real alignment with company values such as teamwork and continuous learning (you may include the company values here).”

You can add:
“Organize the questions into clear pillars (technical, behavioral, situational), explain the purpose of each question, and describe what indicates a strong vs weak answer.”

Essential Notes

• If AI gives repeated or weak questions, train it progressively: “These are level 1 questions, I need level 2,” then “level 3,” and so on.
• Define your preferred tone and the emotional environment you want for the candidate.
• State your detailed objective:
  “I need 5 well-designed interview questions in a friendly, professional tone that makes the candidate feel comfortable and confident, aimed at opening a conversation where they can demonstrate real experience in analyzing complex financial data and handling real-world situations.”
• AI can suggest questions, but selecting the best ones remains your responsibility. Adapt wording using your judgment and knowledge of local candidate behavior.

8) Building Interview Evaluation Criteria Using AI

This step helps you convert professional judgment into a structured written framework, making candidate evaluation consistent and reducing mood-based assessments.

To maximize benefit:

• Define your evaluation goal: best technical performer, best growth potential, or balance.
• Separate must-haves from nice-to-haves.
• Specify the candidate level precisely.
• Avoid generic criteria, request clear behavioral indicators.

Practical Example

“You are an HR specialist at a Saudi financial company undergoing rapid digital transformation. Design interview evaluation criteria for a mid-level Financial Systems Analyst role, aimed at selecting the best technical performance objectively, while considering legal compliance requirements in Saudi Arabia.”

You can add:

• “Divide criteria into clear pillars (technical competency, analytical thinking, communication, culture fit).”
• “Rank criteria by importance and alignment with the job description (attach key points if available).”
• “Assign a suggested weight percentage to each pillar reflecting first-year importance.”
• “Propose a practical scoring system (1–5) with a short explanation for each score.”
• “Provide strong vs warning indicators for each pillar—without absolute judgments.”
• “Add notes about factors that cannot be fully measured numerically.”

Note

You don’t have to create all criteria from scratch using AI. You can ask it to improve your existing criteria and detect bias risks, proposing changes that enhance fairness and consistency. You can also ask for a simplified version usable by HR teams without sacrificing quality.

9) Creating an Onboarding Task Checklist

You might think AI is limited to hiring and scheduling, but it can also support onboarding and integration (and even training). You can use it to create a practical checklist that standardizes onboarding steps. (Talentera’s ATS integrates this capability through its connection with AfterHire.)

Practical Recommendations

When creating an onboarding checklist, clarify:

• When the checklist begins (offer acceptance, contract signing, or before day one).
• The role and seniority level.
• Separate administrative, training, and social integration tasks.
• Define what must be done by timeframe.

Practical Example

“You are an HR specialist at a mid-sized Saudi financial company preparing to onboard a new Digital Marketer. Design a comprehensive onboarding checklist from offer signing through the first 90 days. Divide it into phases (pre-joining, first week, first month, first 90 days), and consider compliance requirements in Saudi Arabia.”

You can improve results by adding:

• “Assign the owner for each step.”
• “Highlight critical points that may harm the employee experience if missed.”
• “Provide a short daily-reference version.”
• “Add simple indicators to measure onboarding success within 90 days.”

Final Guidelines for Professional AI Prompts

These tips make AI prompting easier in recruitment, but the process only matures through testing and comparison. Consider the following:

• Align internally (HR and leadership) on what AI will be used for, and where it will not be used.
• Train AI gradually and introduce your preferences and instructions over time.
• Attach documents or files that enrich context and improve results.
• Ask AI to self-evaluate its output and use its improvement notes (e.g., paste its response and ask: “What are the weaknesses in this job description, and how would you improve it?”).
• Verify important information, ensure it is valid, and confirm sources.
• Judge quality yourself and ensure the output matches your employer brand voice and culture.

Finally, it is extremely important not to share any sensitive information about your company, employees, or candidates with third parties or untrusted platforms that do not clearly disclose how they protect your data.