REC.01
CV Screening with Bias Audit
CV screening that quotes evidence against written requirements instead of scoring, ranking or rejecting anyone, audited for extraction bias before launch. A recruiter decides who moves forward.
Open pageIndustry
AI in recruitment.
A recruiter decides.
The admin is automated. The decision is not. Screening evidence, sourcing, outreach and scheduling, built for a sector the EU AI Act treats as high-risk. The administrative load is automated; every decision about a candidate stays with a person.
Last reviewed 30 October 2026 · Engineering guidance, not legal or regulatory advice.
In context
AI in recruitment means software that handles the administrative volume of hiring (reading applications, drafting outreach, booking interviews) so recruiters can spend their time on candidates. Under the EU AI Act, AI used to filter or evaluate candidates is high-risk, so a person must make, and be able to explain, every decision.
Where it breaks today
01
Volume without signal
Generative tools made applying almost free, so a role that drew fifty applications now draws several hundred, many tailored by AI to the advert. Keyword filters in the ATS reward whoever copied the advert best, which is the opposite of what a recruiter wants.
02
Black-box ranking is a legal risk
The EU AI Act lists AI used to filter applications or evaluate candidates as high-risk (Annex III), with obligations on oversight, logging and transparency. GDPR Article 22 restricts decisions based solely on automated processing. A score nobody can explain fails both.
03
Admin eats the relationship
Chasing diaries, rescheduling, sending the same update to forty candidates, logging notes in Bullhorn or Greenhouse. The work that makes a good recruiter, talking to people, is what gets squeezed.
OUR APPROACH
Evidence for the recruiter, not a verdict on the candidate
The systems we build in recruitment never rank, score or reject a candidate. They do something narrower and more defensible: they read an application against the requirements the hiring manager wrote down, and show the recruiter where each requirement is or is not evidenced, quoting the CV. The recruiter decides who moves forward, and the record shows what they saw when they did.
That design is deliberate. It keeps the decision with a person who can explain it, makes the criteria visible so they can be challenged, and makes bias measurable: we audit whether the extraction finds the same evidence equally well across groups, and the firm monitors outcomes. Whether a particular system falls within the AI Act’s high-risk category, and what that requires of you as deployer, is your determination; we build so that the obligations can be met.
HOW IT DEPLOYS
PHASE 01
Map the desk
We follow real requisitions through your ATS and diaries to find where recruiter hours go, and which steps touch a decision about a candidate.
PHASE 02
Automate the admin
Scheduling, updates and notes first: high volume, no judgement about the candidate, and quick to measure.
PHASE 03
Add evidence, with an audit
Screening support only after a bias audit on your own historical applications, with oversight, logging and candidate notices designed in from the start.
What we build
01
CV screening with bias audit
AI resume screening that extracts evidence against stated requirements, never a score, with an extraction audit across groups before go-live and outcome monitoring after.
SEE CV SCREENING →02
Candidate sourcing and outreach
Searches across your own database and lawful sources, and first-draft outreach a recruiter edits and sends, so the message sounds like a person who read the profile.
SEE CANDIDATE OUTREACH →03
Interview scheduling
Panels, time zones, rooms and reschedules handled against real calendars, with candidates able to pick and change slots without an email chain.
SEE INTERVIEW SCHEDULING →04
Onboarding handoffs
Once a person has made the offer: contracts, right-to-work document collection and first-day logistics passed cleanly from the ATS to HR systems.
SEE HR ONBOARDING →Controls we implement
NO AUTOMATED REJECTION
BIAS AUDIT BEFORE GO-LIVE
CRITERIA VISIBLE AND VERSIONED
CANDIDATE NOTICE BY DESIGN
DECISION LOG PER CANDIDATE
GDPR-ALIGNED RETENTION
These are engineering controls we build into the architecture, not certifications. QQuantum.ai holds no certifications, and the AI Act classification, conformity steps and deployer obligations remain yours — see our EU AI Act notes.
Recruitment use cases
-
-
REC.02
Candidate Outreach
Semantic search over your own candidate database first, and outreach drafts that cite the profile, edited and sent by a recruiter, with GDPR notices and opt-outs built in.
Open page -
REC.03
Interview Scheduling
Interview scheduling across panels, time zones and reschedules, from live Google or Microsoft calendars, written back to the ATS. It arranges meetings a recruiter decided on; it never judges candidates.
Open page
Questions
Questions about Recruitment & Staffing.
How is AI used in recruitment?
Mostly for administrative volume: reading and summarising applications, searching candidate databases, drafting outreach, scheduling interviews and logging notes in the ATS. The contested use is screening, where AI filters or ranks candidates. That use is high-risk under the EU AI Act, and the systems we build keep it to evidence extraction, with a recruiter making and recording every decision.
Will recruitment jobs be replaced by AI?
The administrative share of the job shrinks: scheduling, chasing and first-pass reading. The parts that decide whether a placement works, understanding what a hiring manager actually needs, judging a person and persuading them to move, do not automate well and are where the law expects a human anyway. Recruiters who hand the admin to software get more of the job that matters.
Is AI in hiring legal in the EU?
Yes, with obligations. The EU AI Act classes AI used for recruitment and selection, including targeted job adverts and filtering or evaluating applications, as high-risk, which brings requirements on human oversight, logging, transparency to candidates and data quality. Emotion recognition in the workplace is prohibited outright. GDPR also limits decisions based solely on automated processing. Your legal team determines how each applies.
Is an AI interview a red flag for candidates?
It can be, if it means a system is judging the candidate without a person involved. We do not build systems that assess candidates in interviews or analyse their faces, voices or emotions; emotion recognition at work is banned under the AI Act. Using software to schedule an interview, or to transcribe it for the interviewer with consent, is a different and much more defensible thing.
Which applicant tracking systems do you integrate with?
The ones our clients run: Bullhorn, Vincere and JobAdder on the agency side, and Greenhouse, Lever, Workday, SmartRecruiters and Teamtailor in-house, through their official APIs. The ATS stays the system of record, so every note, status change and decision the recruiter makes is stored where your team already looks.
