How Woolworths Screens 1 Million Candidates a Year With AI — 83% Faster and With a 30% Gender Diversity Gain
By Tim Kreling, Co-Founder, OVI
Candidates rate their AI interview experience 9.2 out of 10. That is not a typo, and it is not a cherry-picked metric from a startup pilot. It is the satisfaction score from Woolworths Australia — the country's largest private employer — after replacing every first-round phone screen in its annual pipeline of one million candidates with Sapia.ai's chat-based AI interview platform.
The assumption that AI screening hurts candidates turns out to be exactly backwards. When Woolworths deployed Sapia.ai, candidates did not just tolerate the new process — they preferred it. And the organisation got something it was not expecting: independent university research would later show that the same system was closing gender gaps by 30%.
The High-Volume Challenge
Woolworths commands 37.4% of Australia's grocery retail market and employs more than 200,000 people. Each year, the company fills roughly 40,000 roles drawn from a pool of one million applicants. Eighty-seven percent of those positions are casual, and 60% of all hires are concentrated between August and December during a seasonal surge that stress-tests every recruitment function in the organisation (Source 1).
When COVID-19 hit, Woolworths had to hire 27,000 new workers in under ten weeks — a scale that would have been impossible with traditional phone screens (Source 1).
The old model was breaking. Phone screens at this volume consumed thousands of recruiter hours and produced completion rates hovering around the industry average of roughly 50%. Woolworths needed a process that could absorb million-candidate throughput without sacrificing the candidate experience that a consumer-facing brand depends on.
The Deployment: Two-Stage AI Screening
Woolworths deployed Sapia.ai's two-stage process integrated with its SAP SuccessFactors ATS (Source 1; Source 5):
Stage 1 — Smart Interviewer (chat-based pre-screen). Candidates receive a link from the job posting and complete a structured text-based conversation on their phone or computer at their own convenience. Organisational psychologists designed five core questions per role type to evaluate soft skills, behavioural traits, and cognitive ability. There is no scheduling, no phone tag, and no time pressure. The system analyses responses and produces a ranked shortlist for hiring managers (Source 5).
Stage 2 — Asynchronous video interview. Shortlisted candidates record video responses on mobile at a time that suits them. There is no live interviewer and no scheduling friction.
Every candidate — successful or not — receives personalised feedback on their performance, a feature that turned the screening process into a brand-building touchpoint (Source 1).
The Outcomes: Speed, Satisfaction, and Scale
The results were immediate and measurable:
- 83% reduction in time-to-hire. In some cases, Woolworths moved from job advertisement to offer letter in 24 hours (Source 1).
- 5,000 recruitment hours saved in the first week alone after deployment (Source 1).
- 82.6% chat interview completion rate — well above the roughly 50% industry average for phone screens (Source 1).
- 9.2/10 candidate satisfaction for the chat interview, 9/10 for the video interview, and an 8.8 NPS (Source 1).
- Nearly 50% of interviews completed on mobile, reflecting the accessibility of an asynchronous, no-scheduling process (Source 1).
As Keri Foti, Woolworths' HO Advisory & Talent Acquisition Services lead, put it: "We love the tool... we could measure advocacy for Woolworths at the end of it" (Source 1).
The Diversity Revelation
Speed and satisfaction were the expected wins. The diversity outcome was not.
In 2023, independent research by Monash University and the University of Gothenburg examined what happens when organisations use AI chat-based screening in traditionally male-dominated roles. The findings were striking: a 30–36% reduction in the gender gap across the full hiring process (Source 4).
The mechanism was twofold. First, when candidates learned that AI — rather than human evaluators — would assess their applications, 30% more women applied. Female candidates perceived the AI-assessed process as fairer than human evaluation, based on their experience with bias in human-led hiring (Source 3; Source 4).
Second, the AI score functioned as an objective anchor that neutralised gender bias at the evaluation stage. When human evaluators were shown applicant gender alone, more than 60% selected male candidates. When shown gender alongside the AI score, selection reached equal representation (Source 4).
Professor Andreas Leibbrandt, who led the research at Monash, said: "AI is treated with apprehension, particularly in recruitment, but these findings show that a transparent and ethical AI system removes many of the barriers faced by women" (Source 3).
Woolworths is the flagship case for this approach at scale. Qantas and Suncorp also use the same Sapia.ai platform, but Woolworths' one-million-candidate pipeline is the largest AI-screened hiring operation documented to date (Source 1; Source 3).
What HR Leaders Should Take Away
Woolworths' deployment challenges a persistent assumption: that AI screening trades candidate experience for efficiency. The data shows the opposite. Candidates prefer asynchronous, low-pressure chat interviews to phone screens, completion rates nearly double, and — when the AI system is designed to evaluate skills and behaviour rather than demographic markers — diversity outcomes improve measurably.
The critical design requirements matter. For AI to reduce rather than amplify bias, systems must exclude demographic data from scoring models, undergo continuous bias testing, and provide transparent results (Source 4). Woolworths' results are not automatic — they reflect deliberate system design choices.
What is Sapia.ai and how does it work?
Sapia.ai is an AI-powered interview platform using a text-based Smart Interviewer chatbot for first-round screening, followed by asynchronous video interview.
How does AI screening improve diversity?
Monash/Gothenburg research found AI chat screening reduced gender gap by 30-36%: more women apply, and AI scores anchor against human bias.
What completion rates does AI chat screening achieve?
Woolworths achieved 82.6% completion rate vs ~50% industry average for phone screens.
Can this scale apply outside retail?
Yes. Qantas and Suncorp use the same platform. The process is industry-agnostic.