Job applicants suing Workday have asked a federal court to certify a class. The case, Mobley v. Workday, is in the Northern District of California. It claims Workday’s AI hiring tools screened applicants out on the basis of race, sex, age, and disability. Judge Rita Lin has set the class certification hearing for March 9, 2027. Workday denies that its technology discriminates, and no court has found that it does.
The proposed class covers applicants who were screened, ranked, scored, shortlisted, or rejected through Workday’s AI tools. The motion names HiredScore Spotlight, Fetch, and Candidate Skills Match. It asks for subclasses by race, gender, age, and disability. The claims are for disparate impact under Title VII, the Americans with Disabilities Act, and California’s Fair Employment and Housing Act, which means the plaintiffs don’t have to prove anyone intended to discriminate.
The motion alleges rejection disparities for Black applicants of more than 18 standard deviations from expected outcomes, and more than 15 for women. It also says many rejections arrived soon after people applied, often in the evening, overnight, or on weekends, which the plaintiffs argue points to automated decisions. These are allegations, not findings.
Workday argues that the plaintiffs lump together different products and hiring situations. It says outcomes depend on individual factors: an applicant’s qualifications, the job’s requirements, the employer, and the decisions each customer makes.
The case got here in steps. In 2024, the court ruled that Workday could plausibly be liable as an agent of the employers using its software. In May 2025, it preliminarily certified a narrower age-discrimination group. In June 2026, it granted part of Workday’s motion to dismiss and denied the rest.
The question in this case isn’t only whether the tool was biased. It’s who answers for it.
Why this matters to employers that buy hiring tools
If a vendor can be treated as the employer’s agent, both can end up in the case. Workday’s own defense, that outcomes depend on each customer’s choices, points back at the employer’s configuration and decisions. Either way, an employer can’t assume the liability sits with the vendor.
Questions to ask about your hiring tools
- Which steps in your hiring process does an AI tool screen, rank, score, or reject on its own, and which settings did you choose?
- Has the vendor tested for adverse impact by race, sex, age, and disability, and will it share the results?
- Do you run your own analysis of hiring outcomes, or rely only on the vendor’s?
- How quickly do rejections go out, and does a person see any of them first?
- What does the contract say about responsibility if a tool is found to discriminate?
- Can an applicant ask for a person to review the decision?
Buying a hiring tool doesn’t buy you out of the hiring decision.
Sources