AI in Talent Acquisition: Automation With Humans in the Loop

AI in talent acquisition works best when AI agents handle the repetitive recruitment stages and people make the important hiring decisions. Agents can screen resumes, contact shortlisted candidates, schedule and conduct first-round interviews, and monitor those interviews to flag fraud attempts. But recruitment teams set the hiring criteria and review the AI recommendations, and hiring managers decide who gets the offer.

This blog covers why human review matters, how our agentic recruitment software, Berri 360, automates the recruitment process while recruitment teams still make the decisions, and the compliance and security questions to ask before a pilot.

Why Should Humans Review AI Recommendations in Talent Acquisition?

AI agents can run recruitment stages at scale, but the decisions that affect a candidate’s career need human review. A rescheduled interview does little harm. But mistakenly rejecting a qualified candidate costs the company a good hire and costs the candidate a fair chance at the job.

Recruitment teams need to review AI recommendations for the following three reasons:

1. Hiring Managers Make the Final Hiring Decision

In the first round, an AI interviewer assesses every candidate in depth on the same criteria and then recommends the top candidates based on their interview performance. From there, hiring managers may run a second-round interview or assess how a candidate fits the team, because they know the team’s needs and are accountable for the outcome.

2. Some Candidates Trust AI Less Than They Trust People

A 2025 Gartner survey of 2,918 job candidates found that only 26% trust AI to evaluate them fairly. Employers who tell candidates that recruiters review every AI result and hiring managers make the final decision give them an honest answer to that concern.

3. Interview Fraud Decisions Need Evidence a Person Can Stand Behind

An EY India study of background checks found that experienced professionals accounted for 79% of employment fraud cases in IT/ITeS. AI can monitor every interview for proxy interviews and impersonation and flag each attempt with timestamped evidence. Rejecting a candidate for fraud is a serious step, and candidates may challenge it. When a recruiter has reviewed the flagged moments and the interview recording, the company can show that evidence as proof and stand behind its decision.

Each of these reasons points to the same design: AI does the volume work, and people make the decisions. Berri 360 is built this way.

How Does Berri 360 Automate Talent Acquisition While Recruitment Teams Make the Decisions?

Berri 360 is agentic recruitment software that automates recruitment from resume screening to L1 interviews and handles candidate communication through onboarding. Its four AI agents work together and hand candidates from one stage to the next, without someone from the recruitment team having to step in at every handoff. Recruitment teams choose each hiring campaign’s stages and thresholds and can override any result, while hiring managers make the final hiring decisions. Here is how the AI agents handle each recruitment stage:

1. Berri Search & Match

Berri Search & Match screens resumes contextually against the job description, looking at each resume’s overall context rather than matching keywords alone, and scores every candidate from 0 to 100. Recruiters set the pass mark for each campaign, can view the skills identified in each resume, and can override any result they disagree with.

2. Berri Connect

Berri Connect reaches shortlisted candidates by voice call, WhatsApp, SMS and email, checks their interest, schedules interviews and answers their questions 24/7. If a candidate misses a call, Berri Connect can send a follow-up message, and candidates can request a call back. Recruiters can also review a record of every call and chat Berri Connect has with each candidate.

3. Berri MasterMind

Berri MasterMind conducts L1 interviews and scores every candidate on technical, communication and behavioral skills against the same written rubric for the role. It can run more than 20,000 interviews a day, which helps in campus hiring and other high-volume drives where interviewer fatigue is a problem. Hiring managers review the interview reports and decide who moves to the next round.

4. Berri Proctor

Berri Proctor monitors interviews in real time for impersonation and proxy attempts and flags each attempt with timestamped screenshot evidence. Recruitment teams review that evidence before any adverse decision affects the candidate.

The four agents plan and coordinate this work on one shared candidate record. Berri 360 natively integrates with Workday and SAP SuccessFactors, so candidate data from every stage updates in the ATS automatically and nobody has to re-enter data or chase handoffs. Every move of a candidate between stages is also logged with the name of the person who made it, which gives recruitment teams an audit trail of hiring decisions.

Companies automating recruitment with Berribot report these results:

• 9.5 recruiter hours saved per offer

• Time to offer cut from more than 25 days to 10 days

• 78% savings in recruiter and panel hours

• 2x improvement in candidate show-up rates

Faster hiring also has to hold up to legal and security review, which is where human oversight becomes a requirement rather than a preference.

What Compliance and Security Rules Apply to Using AI in Talent Acquisition?

AI regulations increasingly require human oversight when AI is used in talent acquisition, and data protection laws cover the candidate data it processes. 

Under the EU AI Act, AI used for recruitment and candidate selection is classed as high-risk, with obligations that include human oversight applying from December 2, 2027. The Act has also banned emotion recognition in the workplace, including in recruitment, since February 2025. In India, the Digital Personal Data Protection (DPDP) Rules bring core obligations on notice, consent and security safeguards into force in May 2027, and these cover candidate data too.

Berribot’s safeguards follow the same principles. Berribot holds SOC 2 Type II, ISO 27001 and India DPDP Act certifications and supports GDPR and CCPA/CPRA requirements, with details available in the Trust Center. Candidate data is encrypted at rest and in transit, access is role-based, and each company’s records are kept separate. 

Biometric data is processed only with explicit consent or a documented employment-law basis, and candidate data is kept for 30 days after the recruitment activity ends by default, as set out in the data retention policy. Berribot’s AI does not infer emotion, personality or mood from a candidate’s face or voice, and recruitment teams review results before any adverse decision.

Whichever AI recruitment platform you evaluate, ask the vendor:

1. Which recruitment stages does the AI act on alone, and which need a person’s approval?

2. Can recruiters see why a candidate was ranked or scored a certain way, and override the result?

3. Who reviews an interview fraud flag, and what evidence do they see?

4. Is every change to a candidate’s status logged with the name of the person who made it?

5. How long is candidate data kept, and how is consent collected for face and voice checks?

Ask us all five questions. In a 30-minute demo, we’ll answer each one on a live role.

Book a demo

Frequently Asked Questions

1. What Is Human-in-the-Loop Recruitment?

Human-in-the-loop recruitment means AI agents carry out defined recruiting tasks, such as resume screening, candidate outreach and L1 interviews, while humans make the decisions that affect a candidate’s outcome, such as shortlisting, rejecting or making an offer.

2. Does Berri 360 Make Hiring Decisions?

No. Berri 360 automates recruitment stages and produces candidate rankings, interview reports and fraud evidence. Recruitment teams can override any result and move or close any candidate at any time, and hiring managers make the final hiring decisions.

3. Is AI in Recruitment Regulated?

Yes, in a growing number of regions. The EU AI Act classes AI used for recruitment as high-risk, with obligations including human oversight applying from December 2, 2027. In India, the DPDP Rules require notice, consent and security safeguards for personal data, including candidate data, from May 2027.

See Berri 360 on One of Your Roles

Bring one job description, and we’ll show you in 30 minutes how Berri 360 handles the recruitment work while your team makes the decisions.

Book a demo