Imagine you are assigning a task to a new recruiting intern. If you simply say, “find the best,” they will come back with a dozen clarifying questions: “What exactly should I look for in the CV?” or “What kind of experience is considered sufficient?”
An AI assistant works the same way, but it cannot ask follow-up questions. Its effectiveness depends 100% on how detailed your candidate persona is "written" into the system.
AI excels at verifying concrete information. Let automation handle the routine checks of basic requirements (experience, certificates, tool proficiency), while you save your time for evaluating soft skills during the interview.
Do not mix multiple skills into a single point. If a candidate possesses only one of them, the AI may produce an incorrect result.
❌ Avoid
“Experience in SMM and email marketing”
✅ Try
1. “Experience managing corporate social media accounts”
2. “Experience working with email marketing services”
Words like "expert," "strong," or "extensive experience" are interpreted differently by everyone. To make the AI work accurately, translate these concepts into measurable metrics.
Add timeframes: instead of "experienced," write "3+ years of experience."
Specify the stack: instead of "knowledge of frameworks," write "experience with React or Angular."
Use quantitative indicators: instead of "managed a large team," write "managed a team of 10+ people."
Examples:
❌ Poor: “Has leadership qualities”
✅ Better: “Has at least 1 year of experience as a Team Lead”
❌ Poor: “Knows English”
✅ Better: “Listed English level is Upper-Intermediate (B2) or higher”
You can formulate criteria to define what a candidate must have or, conversely, what they should not have.
✅Positive criterion example:
"Candidate has experience leading B2B sales departments in the SaaS sector."
✅Negative criterion examples:
"Candidate does not reside in Ukraine."
"Most recent experience is not as a freelancer."
The system only analyzes what is in the text: the CV and the cover letter.
✅Works perfectly
"Proficient in Python at a Middle+ level"
"Holds an AWS Certified certificate"
"2+ years of experience in B2B sales"
❌Save for the Interview
"Has an analytical mindset"
"High potential for learning"
"Shares our company values"
AI speeds up the process by showing how well a candidate matches your requests, but it does not make the final choice for you.
Prioritization, not disqualification
Use AI scores to understand whom to review first. Don't automatically reject candidates just because of a low score.
Context check
Always open the candidate's profile. AI might miss a nuance that you, as an expert, will spot at a glance.
Right to veto
You can always manually change the AI's rating. Your intuition and experience take priority over algorithms.
