AI in hiring deserves scrutiny - discrimination risk is real, and hand-waving isn't an answer. Here is exactly how Lemonade's screening is built to be fairer than the process it replaces.
Every candidate for a role is screened against the same brief - the role's actual requirements, checked one by one against the same criteria - which decades of hiring research shows is more consistent and less biased than unstructured CV sifting.
Every score in every screening report links to the evidence behind it - what's actually in the candidate's record and application. Nothing is a black-box rating, and gaps are flagged honestly rather than hidden.
Lem does the legwork - sourcing, screening, the paperwork - and every verdict is shown with its reasoning. People Lem finds and approaches are never turned down by software: Lem recommends, and you decide. When someone applies to a role advertised through Lemonade, their application is screened automatically, and it can be turned down without a person if it misses one of the role's fixed requirements (for example the right to work in the UK, or a question you marked as essential) or scores below the level we put forward. They are told this before they apply. If it happens, the email says the decision was automatic and which check said no, and they can reply to ask a person at Lemonade to look again. That person replies within 5 working days, can change the result, and tells them what they found either way.
Screening is competency-based: your role's actual requirements, checked one by one, the same way for every candidate, with the evidence shown. Nothing about a person that is not a requirement of the role is scored.
UK GDPR compliant, ICO registered (ZC092497). Candidates are never shared without consent - nobody appears on a shortlist without saying yes to the role - and candidate data is never sold.
Questions about any of this? Ask us before you post a role - we'd rather you were sure. See also our Privacy Policy.