Swyg is on a mission to make recruitment more fair for everyone. We believe that hiring the best people in the world will naturally lead to diverse teams. But to find who is the best candidate for a job, we all have to put aside old assumptions.
That’s why we’ve built our platform on a solid scientific foundation. There is no silver bullet, but if we consistently enforce best-practices at every stage, we can make a difference. That’s why we designed a process, from the ground up, to be less biased.
In a nutshell, every candidate does multiple 1:1 interviews with other candidates while alternating between the role of interviewer and interviewee.
Let’s get into detail about how that process helps reduce bias.
1. No reliance on static profiles
Because we make face-to-face interviews accessible to more people, we can reduce reliance on static profiles (LinkedIn/CV/Resume) which are a major source of bias. We shift the emphasis from past job titles to actual demonstrable skills which we assess through interviews.
2. Multiple interviewers per candidate
Every candidate is interviewed by multiple interviewers. This is a really powerful concept. It means that no one individual can sway the outcome. Whether they are biassed, tired, or just having a bad day.
Having multiple interviewers means we get a more representative range of views about a candidate, but more importantly, it means we can detect and eliminate reviews that differ from the consensus. We can also identify cases where no consensus exists so more data is needed.
3. Each interviewer interviews multiple candidates
This tells us if a reviewer is reliable. Are their reviews always in line with those of other reliable reviewers, or are they putting random reviews? We can easily see this by seeing how one individual rates multiple candidates. We also use this data to generate interviewer profiles.
4. An interviewer pool that’s representative of the candidate pool
Affinity bias, is really common. Companies naturally tend to build monocultures because the selection process is managed by the same people that were already hired by the company. We solve this problem by using, by design, a pool of interviewers that is representative of the pool of candidates.
5. AI-Generated Interviewer Profiles
This is the really cool part of our platform. You can imagine that if you had a pool of candidates that you know really well, you could use them to calibrate your interviewers. Similarly, if you had a group of interviewers that you understand very well (in terms of how they rate, their biases, etc) you could use them to get to know your candidates better. What our AI does is to solve both these problems simultaneously in real time by having peers alternate between the role of interviewer and interviewee.
Just the Science
To avoid confusion: Swyg is not a racism detector. We do not have magical AI that can tell what you’re thinking by the way you raise your left eyebrow before you ask a question. We just follow the science: We build best-practices into our platform from the ground up and use AI to super-charge those best practices.