Zappyhire

Company overview

Zappyhire is an AI recruitment automation platform founded in 2018 by Jyothis K S and Deepu Xavier and headquartered in Kochi (Ernakulam), Kerala, India, with a secondary office in Maharashtra. It markets an 'agentic AI' suite to medium and large enterprises spanning intelligent resume parsing, a recruiting chatbot, gamified and AI assessments, a 'robotic' (automated) video interview that analyzes body language and speech, automated candidate scoring and ranking, and an underlying applicant tracking system. Its core differentiator is automated pre-screening and shortlisting at scale; public pricing starts around USD 39 per user per month, placing it in the mid tier with enterprise deals on request.

Regulatory exposure

Zappyhire's automated video interviewing, candidate scoring, and ranking place it squarely in the EU AI Act Annex III 'high-risk' category for recruitment AI and within scope of NYC Local Law 144's AEDT bias-audit regime where it is used by NYC employers. Despite this, the vendor publishes no bias audit, no model card or Article 11 technical documentation, no Fundamental Rights Impact Assessment material, and no EU AI Act deployer guidance. It markets enterprise-grade security (ISO 27001, SOC Type 1 and 2, GDPR, VAPT logos) and a GDPR page describing its processor role and Article 30 audit logging, but its product pages repeat an unevidenced 'eliminate unconscious bias' claim that was singled out in peer-reviewed academic criticism (Drage and Mackereth, Philosophy & Technology, 2022), creating notable exposure given that the AI scoring methodology and demographic-impact testing are entirely undocumented.

Path to a higher score

The fastest gains would come from commissioning and publicly posting an independent bias audit (e.g., a NYC LL 144-style disparate-impact analysis) and publishing a model card / explainability statement describing how the video-interview and scoring models work and what data they exclude. Standing up a public trust center with downloadable ISO 27001 and SOC 2 reports (rather than logos), an EU AI Act / FRIA deployer-obligations pack, a public status or security-disclosure channel, and an automated-decision-making disclosure in the privacy policy would move every rubric category upward and substantiate the bias claims currently asserted without evidence.