Unilever's AI Hiring: 90% Faster, 16% More Diverse

The Challenge

Unilever, a consumer goods giant handling 1.8 million job applications annually, struggled with a manual recruitment process that was extremely time-consuming and inefficient [1]. Traditional methods took up to four months to fill positions, overburdening recruiters and delaying talent acquisition across its global operations [2].

The process also risked unconscious biases in CV screening and interviews, limiting workforce diversity and potentially overlooking qualified candidates from underrepresented groups [3]. High volumes made it impossible to assess every applicant thoroughly, leading to high costs estimated at millions annually and inconsistent hiring quality [4]. Unilever needed a scalable, fair system to streamline early-stage screening while maintaining psychometric rigor.

Quantitative Results

• Time-to-hire: **90% reduction** (4 months to 4 weeks)
• Recruiter time saved: **50,000 hours**
• Annual cost savings: **£1 million**
• Diversity hires increase: **16%** (incl. neuro-atypical candidates)
• Candidates shortlisted for humans: **90% reduction**
• Applications processed: **1.8 million/year**

Implementation Details

Technology Partners and AI Stack

Unilever collaborated with Pymetrics (acquired by Harver in 2023) for its neuroscience games, which use machine learning to benchmark candidates against successful employee profiles across 50+ traits like risk-taking and attention [2]. Video interviews were powered by AI platforms like HireVue, employing computer vision for micro-expressions, natural language processing (NLP) for speech patterns, and psychometric models for personality insights [1]. The stack integrated seamlessly with Unilever's ATS for end-to-end automation.

Pilot and Rollout Timeline

Implementation began in 2016 with a pilot for US Future Leaders Programme, processing 250,000 applicants in months [3]. By 2018, it expanded to all entry-level hires globally, reaching 1.8 million apps/year by 2019 [2]. Full rollout across 70+ countries occurred by 2020, with continuous model retraining on diverse data to combat bias. Challenges like data privacy (GDPR compliance) were addressed via ethical AI frameworks and third-party audits [5].

Process Workflow and Bias Mitigation

Candidates submit CVs screened by AI for keywords and fit. Top applicants play 12 mobile games (15-30 mins) measuring traits non-verbally, avoiding cultural biases [4]. Invited to video interviews (5 questions, recorded), AI scores on energy levels, eye contact, and language calibrated against top performers. Scores generate a 'match index'; only high scorers advance, reducing human assessments by 90% [1]. To overcome initial bias concerns, Unilever used diverse training datasets from 100,000+ past hires and partnered for transparency audits.

Integration and Scalability Challenges

Key hurdles included candidate experience (gamification boosted engagement 20%) and technical integration with legacy HR systems, solved via APIs and cloud deployment [3]. Legal challenges around AI fairness were met with regular algorithmic audits and human oversight for final decisions. By 2025, the system supports hybrid remote hiring post-COVID, with ML models updated quarterly for evolving job needs [5]. Training involved 500+ recruiters on interpreting AI insights, ensuring buy-in.

Current Status and Evolutions

As of 2025, the platform remains core to Unilever's HR, now enhanced with generative AI for personalized feedback to rejected candidates. It handles volume spikes efficiently, with ongoing expansions to mid-senior roles [4]. Metrics tracking via dashboards show sustained ROI, positioning Unilever as an HR tech leader.