AI in recruiting: faster work, human decisions
Artificial intelligence takes repetitive reading and sorting off recruiters’ hands, but a person has to decide about candidates. This is how HRAtan uses it.
What does HRAtan use AI for?
- CV analysis: a CV in any language becomes one consistent profile (skills, experience, a short summary).
- Match score: the CV is compared with the requirements of the job ad, with a short rationale.
- Translation: the decision maker can translate a forwarded candidate’s summary into their own language with one button.
- Job ad writer: a draft ad from a few details about the position (Enterprise plan; can be tried once per company on other plans).
What the AI does not do
- It does not decide: screening, inviting and selecting stay with the recruiter; the score only helps with ordering.
- Candidates can object to AI-based assessment; then the recruiter assesses manually.
- We do not train our own AI models on candidate data; analysis is performed by external AI providers as part of the service.
- AI can be switched off for the whole organisation, in which case every candidate is assessed manually.
The AI Act and responsible use
The EU AI Act classifies AI systems used for recruitment and for evaluating candidates as high-risk, and expects human oversight and transparency. That is why in HRAtan the AI’s output is always a suggestion, the decision and its responsibility stay with a person, and candidates are informed about the use of AI. This is information, not legal advice.
Tips for using AI well
- Use the score for ordering, not as a verdict: read the profile and the CV too.
- Write precise requirements in the job ad: the match score measures against them.
- For your favourites, open the original CV before forwarding.
- Plan for the exceptions: candidates who objected and CVs saved as images are assessed manually.
Frequently asked questions
Can AI discriminate?
Any automated assessment can make mistakes. Before analysis, HRAtan removes protected data from the CV (for example date of birth, gender, nationality, marital status, religion, health), and the AI is instructed not to assess them. The filtering is not perfect, so the output is only a suggestion and a person always makes the decision.
Can AI be switched off?
Yes, for the whole organisation. Every candidate is then assessed manually, and all other features keep working.
How many AI analyses do the plans include?
Analyses fall under a monthly quota that depends on the plan, and they can also be tried during the free trial. Current limits are on the Pricing and plans page.