Objectively better candidates from more diverse backgrounds in less time: It’s the veritable holy grail of recruiting. And for years now, recruiters have looked to artificial intelligence (AI) to make that fantasy a reality. Despite the real progress that has been made in this realm — from automated chatbots to candidate-screening algorithms — many cutting-edge tools have fallen short of AI’s promised potential.
That doesn’t mean AI recruiting can’t live up to the hype. It just means we need to approach AI recruiting from a different angle. Upsider.ai, an artificial intelligence sourcing solution that automates candidate identification and engagement to help employers make the best hires, offers a model of how machine learning can really help recruiters tap better, broader candidate pools more quickly.
Keeping the Promise of AI
From the start, recruiters and talent acquisition teams have been drawn to AI because it offered a chance to finally make objective a process that has long relied — not always successfully — on gut feeling. With the right algorithm, recruiters could compare candidates on concrete metrics rather than subjective measures like interview responses, leading to more informed hiring decisions. And this objective assessment opens up incredible possibilities for diversity, equity, and inclusion initiatives. By removing unconscious bias from the equation, AI could make it easier for recruiters and hiring managers to focus on what really matters: candidates’ professional qualifications.
Unfortunately, AI’s application in real-world recruiting has had mixed results. In some instances, it has even worsened the very problems it seeks to solve. Recall when Amazon had to abandon its AI hiring algorithm after discovering it was giving preference to male candidates over women.
Amazon’s AI woes illustrate the tricky problems inherent in bringing AI to the recruiting process. Most legitimate AI-powered hiring solutions are based on a machine learning model. Essentially, the makers of these systems train them to identify best-fit talent by feeding them information about past hires. Because unconscious bias influences our hiring processes, our sets of previously successful candidates aren’t necessarily objective. They may be encoded with all sorts of subtly discriminatory signals, which then skew the algorithm’s results.
That’s exactly what happened to Amazon. Its AI was trained on prior successful Amazon applicants, and most of those applicants were men. Thus, the algorithm unintentionally learned to perceive markers of male gender as markers of a successful candidate.
AI hasn’t been all bad. It has helped plenty of recruiters streamline their processes, source candidates faster, and automate away some of the more mundane tasks of hiring. That’s why so many recruiters are still searching for the perfect AI tool: They’ve seen glimpses of its value, and they want to unlock its full potential. That’s where Upsider.ai comes in.
How Upsider.ai Uses Machine Learning to Diversify Talent Pipelines and Deliver Best-Fit Talent
Upsider.ai is artificial intelligence sourcing software that automates candidate identification and engagement to help employers make the best hires. It is based on a machine learning model that adapts to patterns in users' candidate selection and engagement results, creating a self-evolving sourcing experience that learns and becomes more efficient over time.
It’s worth mentioning that Upsider.ai uses real machine learning models because a significant chunk of companies purporting to use machine learning don’t really use it. “AI” is a marketing buzzword for some recruitment tech companies, but not for Upsider.ai. By paying attention to how recruiters use the platform, Upsider.ai’s machine learning model comes to understand what they’re looking for. It builds an ideal candidate model based on the recruiter’s sourcing behavior, and then it matches the records in Upsider.ai’s database of 120+ million candidates against that model to identify good fits.
How does that compare to other AI sourcing platforms that match candidates to jobs? Most matching engines rely on keyword searches: Recruiters punch in specific job titles, and the system surfaces candidate profiles that contain those keywords. Then, the AI kicks in to match these candidates against the recruiter’s open job.
Upsider.ai operates differently. While users can search its database for strict keyword matches, Upsider.ai also has a unique AI search function that expands a recruiter’s search to encompass job titles that don’t match the keywords but do match the recruiter’s requirements. Upsider.ai also allows recruiters to search for candidates by companies they’ve worked for. As with the title search, Upsider.ai can expand a recruiter’s search to include candidates with experience at similar companies not expressly named by the recruiter.
For example, say you need a back-end developer. A traditional keyword search would only surface candidates who have held the exact title of “back-end developer.” Upsider’s AI search, on the other hand, would also surface full-stack developers, because these developers have back-end experience despite holding a different job title.
That’s a relatively simplified example, but it illustrates the core difference between Upsider.ai and typical candidate-matching tools. Upsider.ai uses machine learning to not just rank candidates, but to surface a richer set of candidates to be ranked in the first place. By bringing these non-obvious matches to recruiters’ attention, Upsider.ai helps companies broaden their talent pipelines to include more kinds of candidates from more diverse backgrounds. And because these candidates are measured against the ideal candidate model built by Upsider.ai, recruiters are only presented with the best fits.
Upsider.ai circumvents the bias that regrettably creeps into other AI-matching algorithms by bringing AI to an earlier stage in the sourcing process. It recognizes things like parallel experience, transferable skills, overlapping responsibilities, and other easy-to-miss qualifications that slip by other AI sourcing solutions. And in doing so, it delivers on AI’s original promise to recruiters: objectively better candidates, from more diverse backgrounds, in less time.