AI Recruitment Calls: How to Identify a Compliant and Legitimate Approach

Millions of candidates today receive automated calls as part of recruitment processes. Most don't know what to make of them: is this legitimate? Does it really count? What if you're not comfortable with it?

Here's how to tell the difference between a serious approach and a questionable one, and what the law now requires of every actor involved.

Why AI calls are becoming the norm in recruitment

Recruitment volumes have exploded in recent years, and HR teams have not kept pace. A recruiter manages on average between 50 and 200 applications per open role. Calling every candidate manually for pre-screening takes several hours a day, with results that often vary depending on the time, fatigue, or the recruiter's workload.

AI voice agents address this constraint directly. Available 24/7, they conduct the same pre-screening interview at the same level of quality, whether on a Monday morning or a Sunday at 10pm. In 2025, Omogen conducted over 330,000 automated calls, of which 61,000 were made outside office hours. This figure illustrates something structural: candidates do not look for work only between 9am and 6pm, and recruiters cannot be available at all hours.

The spread of AI calls is also driven by process economics. The cost of a human pre-screening phone interview ranges from 15 to 40 euros depending on the sector, once time, coordination, and no-shows are factored in. Automation brings this cost to a fraction while increasing coverage. For companies recruiting at volume, the calculation is straightforward. To estimate this impact on your business, use our ROI calculator.

Finally, AI calls enable a standardisation that improves fairness. Every candidate is assessed on the same criteria, with the same questions, in the same order.

The 4 signs of a legitimate AI call

Not all AI calls are equal. Some are conducted transparently and respectfully; others involve opaque practices that already violate current legal obligations. Here are the four concrete markers of a compliant call.

1. Immediate and explicit identification. A legitimate AI call begins with a clear announcement: the agent identifies itself by name, states it is an automated system, and specifies which company it is acting for. This transparency is not optional. A call that presents itself as human without ever clarifying the matter is problematic, both ethically and legally. Omogen's voice agent, Alex, identifies itself within the first few seconds of every call.

2. A link to a real application. A legitimate AI call is always tied to a concrete action on your part: an application submitted, a registration on a platform, a response to a job posting. If you receive a call without having ever expressed interest in the company or role concerned, that is a warning sign.

3. Independent verification. You must be able to verify the existence of the company, the role, and the service provider mentioned during the call. A legitimate call gives you enough information to trace the source on your own.

4. AI Act compliance. Since the progressive rollout of the European regulation on artificial intelligence, AI systems used in recruitment fall under the "high-risk" category. This means specific obligations apply, including the right for the candidate to request human review of any automated recommendation, as well as the ability to refuse the call or end it at any time.

What the AI Act says about automated recruitment calls

The European AI Act, which entered into force in 2024 with progressive application through August 2026, classifies AI systems used in recruitment as high-risk applications. This classification carries real weight: it imposes a binding framework on all operators deploying this type of technology in Europe.

The main obligations fall into three areas. First, transparency: every candidate must be informed that they are interacting with an automated system before the exchange begins.

Second, the right to human review: no decision significantly affecting a candidate (rejection, advancement) may rest solely on automated processing without the possibility of human challenge.

Third, sanctions commensurate with the stakes. Breaches of AI Act obligations for high-risk applications can result in fines of up to 30 million euros or 6% of global turnover.

What Omogen does concretely to be compliant

AI Act compliance is not a label you stick on a website. It translates into very concrete design decisions, visible in every call. Discover the detail of our approach on the features page.

Alex, Omogen's voice agent, systematically identifies itself at the start of each call as an AI agent working for the hiring company. Opt-out is built into the flow: if a candidate does not wish to continue the automated interview, they can end it at any time and ask to be called back by a human.

Every interview results in a consultable evaluation report. In 2025, Omogen generated over 335,000 such reports. These reports form the basis of process traceability.

Omogen is also Deeptech-certified by BPI France and supported by players including Mistral and Gladia.

What to do as a candidate if you receive an AI call

Receiving an AI call for a recruitment interview can be surprising the first time. The approach to take is simple and logical.

Start by noting the name of the AI agent and the company it is calling for. If this information is not given to you spontaneously, ask for it. A compliant system will answer clearly.

Then verify the link to an action of yours. Find the application, job posting, or registration that triggered the call. If you find no link, contact the company directly through its official channels to verify.

During the interview, respond as you would with a human recruiter. Questions typically cover your availability, experience, salary expectations, and location.

At the end of the call, you can ask to consult your evaluation report. This is your right under the AI Act. If the pre-selection decision does not suit you, you have the right to request a human review.

What a recruiter must put in place for a compliant AI calling programme

Deploying an AI voice agent in recruitment without a compliance framework is a risky bet. Here are the foundations of a solid programme.

The first step is system documentation. Before any deployment, you must be able to describe precisely how the agent works, what criteria it assesses, and how it weights responses.

The second step is pre-selection criteria configuration. The filters used by the agent must be defined, documented, and validated by the HR team. At Omogen, these criteria are configurable: experience, sector, training, languages spoken, certifications, location. Each filter has an explicit business justification.

The third step is establishing a human review circuit. Every candidate must be able to challenge an automated decision.

Finally, candidate communication is a lever that is often underestimated. Informing candidates in advance, in the application confirmation email, that an automated interview will take place reduces friction and improves call completion rates.

FAQ

Can an AI voice agent make a recruitment decision on my behalf? No. Under the AI Act, no decision significantly affecting a candidate may rest solely on automated processing. The voice agent assesses and qualifies, but the final decision always rests with a human.

How do I know whether the call I received was linked to a real job offer? Check your application confirmation emails for a message from the company concerned. If you find no link, contact the company directly through its official contact details.

What does a company risk if it uses an AI voice agent without respecting transparency rules? The sanctions provided by the AI Act for breaches in high-risk applications can reach 30 million euros or 6% of global turnover.

Further reading

1. High-Volume Recruitment: Why Voice Pre-Qualification Changes the Rules
2. 24/7: Why Reaching Candidates When They Are Available Changes Everything
3. How Myttra Structures Its Fast-Growth Recruitment with Omogen