Skip to content
Security and the AI Act

Everything your lawyer will ask about, before you decide

Under the AI Act, recruiting is a high-risk area. That is why we build and document the system with those requirements in mind today, and you receive the set of documents before you sign anything.

Four principles

How it is built

A person decides

The virtual recruiter recommends, and after an assessment it only proposes a rejection and gives the reasons. A person approves it.

The candidate knows they are talking to an AI

Before the call they hear that a virtual AI assistant runs it and that the call is recorded. They can ask to speak to a person.

The data is protected

We encrypt it at rest and in transit. Each customer's data is kept separate and access depends on each team member's role.

Every decision leaves a trace

The history shows who approved a decision, when, and what the reasons were.

Feb 2025the prohibitions apply, including the ban on emotion recognition at work
Aug 2026the duty to tell people they are talking to an AI
2 Dec 2027the requirements for high-risk systems in recruiting
Documents

What you receive before you decide

Document For whom
Technical documentation IT, security
Model cards and data cards IT, data protection officer
Risk assessment with a discrimination analysis data protection officer, compliance
Data protection impact assessment (DPIA) data protection officer
Instructions for use in Polish and English you as the deployer
Training programme for supervisors HR, supervisors
Model test report IT, compliance
Procedures: breaches, encryption, backups security
Films

How to check an AI vendor

The films are in Polish, some with English subtitles.

Let's talk about a pilot in your team

30 minutes with Monika, co-founder of SmartyTalent. We will show an example of the virtual recruiter's work and answer your questions about the ATS integration, supervision and documents.

Book a 30-minute demo →