By generating Diversity & Inclusion compliant long-lists that randomly add candidates who have close-enough correlations to the user’s request by measuring these variables against the data in EVA’s algorithmic recommendation module
Basic personal information such as name, age, race, native language and gender can trigger a reviewer’s unconscious bias. As you don’t need this information to assess whether a candidate is great for the job, EVA can remove data from the User Interface
Engineered without Bias
EVA’s ML algorithms do not include data that is inherently biased – such as age, gender, location or nationality
The set of input variables is strictly controlled. There is limited scope for bias by excluding inherently biased data from the Machine Learning model’s data sets.
For example, personal data and age, gender or even postcodes are never used. The latter resolve to 50m radius and introduce bias between more or less affluent locations.
The HR 4.0 Way
EVA’s predictive model is based on candidates’ successes and drop-outs across the recruitment funnel from Longlist to Shortlist (Screen) in the interview process and hiring.
In other terms, EVA tests candidate data against your recruitment history and pipelines to identify a list of the most relevant candidates for any given requirements. Decision data against candidate progressions are used to teach EVA’s predictive algorithms within the system.
Engraining Ethical D&I within a sustainable hiring culture
This ensures teams build invaluable and sustainable hiring practices. EVA learns, tracks, and retains the best-practice hiring patterns and insights.
Over time, EVA will gain an understanding of your job(s) market context.
Whether it recognises the difference between a step up and sideways step in a career or anticipating candidate movements for truly proactive recruitment such as predicting who is likely to move within specific departments, job families or career stages.