HR
AI for job ads and candidate screening
Use AI freely to draft clear, inclusive job ads, interview guides and candidate emails from your own role description. Be far more careful with screening: AI that filters, ranks or evaluates candidates is high-risk under the EU AI Act. Keep people reading applications and making every decision, and check with your legal advisor before any AI touches CVs.
Effort:
Medium
Time saved:
Roughly 1–2 hours saved per job ad; screening stays manual
The problem today
Small businesses write job ads in a hurry, so ads are vague and attract the wrong applicants. Then one person has to read dozens or hundreds of CVs on top of their normal job.
How it works, in plain words
Think of a writing assistant for your hiring paperwork. It never reads CVs or picks people.
What goes in: your own role brief: main tasks, must-have and nice-to-have skills, working conditions and, if you publish it, the salary range.
What the AI does: it drafts the job ad in Greek and English, flags wording that may exclude people, and prepares an interview guide, a scoring sheet and candidate message templates.
What a person checks: you edit every draft, and people read every application, run the interviews, fill in the scores and make every decision.
What comes out: a clear job ad, an interview guide and email templates that you personalise and send.
AI that filters, ranks or evaluates candidates is high-risk under the EU AI Act, so keep it out of screening and check with your legal advisor before any AI touches CVs.
How it could work for businesses like yours
Illustrative examples, not client case studies.
A restaurant group on Naxos hiring seasonal staff
Every spring the owners rewrite ads for waiters, cooks and bar staff in a rush, and the same questions come back from applicants. They would write one brief per role and ask AI to draft clear ads in Greek and English, plus a short FAQ on hours and accommodation. People would read every application and decide who to interview, and they would track how many applicants ask basic questions the ad already answers.
A 12-person electronics e-shop in Agrinio
The shop needs a customer service agent, and the owner has no interview structure. They would use AI to turn the role brief into an interview guide and a scoring sheet tied to the must-have skills, which the two interviewers fill in themselves. A realistic first target could be the same core questions for every candidate.
A 9-person digital agency in Kallithea
Candidates wait weeks for any reply, which hurts the agency's reputation. The agency would use AI only to draft acknowledgement, invitation and rejection templates, then personalise and send them itself. They would aim for every applicant to receive a reply within a set number of days.
How to do it, step by step
Write a short role brief yourself: main tasks, must-have and nice-to-have skills, working conditions and the salary range if you publish it. Everything else builds on this.
Ask AI to draft the job ad in Greek and English from the brief. Ask it to flag gendered or age-related wording and requirements you don't really need, then edit it in your own tone.
Use AI to prepare a structured interview guide and a scoring sheet tied to the must-have skills. Interviewers fill in the scores, not the AI.
Use AI for candidate message templates: acknowledgement, interview invitation, polite rejection. Personalise and send them yourself.
Do not use AI to filter, rank or score applications without legal advice. If you consider an applicant tracking system with AI screening, ask the vendor for its AI Act documentation, involve your legal advisor and carry out a DPIA first.
Record how each decision was made: who read which applications, which criteria applied and why. This protects candidates and your business.
Tools to consider
ChatGPT, Claude or Microsoft 365 Copilot: good for drafting job ads, interview guides and candidate emails from your brief. Caveat: do not paste CVs or candidate data into them for screening.
Workable: applicant tracking system with AI help for writing job descriptions and finding candidates. Caveat: check which AI matching or screening features are switched on and what the vendor documents about AI Act compliance.
LinkedIn: AI-assisted drafting when you post a job. Caveat: the AI Act's high-risk list also covers AI that places targeted job ads, so ask how targeting and matching work before relying on them.
Data protection and GDPR
Recruitment is one of the most regulated areas for AI. Under Annex III of the EU AI Act, AI used to place targeted job ads, filter applications or evaluate candidates is high-risk, with duties such as human oversight, record-keeping and informing candidates and staff. These obligations phase in through 2026–2027. Emotion recognition in the workplace, such as reading emotions from video interviews, is already banned. Under the GDPR, candidates have rights around decisions based solely on automated processing, and screening at scale usually needs a DPIA. Check with your DPO or legal advisor before any AI touches applications.
Mistakes to avoid
Pasting CVs into a general chatbot and asking it to rank candidates. This can bring you under the AI Act's high-risk rules, with heavier duties than you expect.
Keeping AI-suggested requirements you don't actually need, which shrinks and skews your applicant pool.
Using tools that claim to read personality or emotions from video or voice.