How machine learning can tackle key recruiting problems

4 mins, 40 secs read time

Recruiters and other talent professionals know there are benefits of staying on the cutting edge, but the innovative and game-changing tech out there today also comes with risks. For example, artificial intelligence and machine learning can help automate manual tasks in your recruiting process, but it can potentially create bias towards candidates and impact your hiring decisions.

In a revealing session at Greenhouse Open 2022, Henry Tsai (VP of Product at Greenhouse) moderated a panel with Susanna Vogel (HR Reporter at Morning Brew), Mona Khalil (Data Science Manager at Greenhouse) and Dave Dyer (Senior Data Scientist at Greenhouse) to discuss how machine learning can be used help tackle recruiting issues and to share Greenhouse’s take on the potential benefits, risks and approaches of doing so. Catch the highlights from their important conversation here.

Key hiring problems that AI and ML can help solve

In this session, Henry highlights four areas where some of the thorniest hiring problems exist and asks the panel how artificial intelligence and machine learning can be harnessed to help talent professionals overcome them.

The first area is finding qualified candidates, posing the question: Are there more efficient ways or better places to find great talent? He outlines the sub-buckets as problems with sourcing diversity, pipeline parity and how candidates – past, present and future – are stored .

The second area is improving and iterating the hiring process. How do you know when you’re doing well in hiring or if you’re getting better? How can machines help us optimize that experience? The sub-buckets in this area include benchmark data, insights to action, setting effective metrics and growing in your roles.

The third area is effective hiring. What does it mean to successfully hire at scale? How can you retain talent? Can you measure the productivity of a team? The sub-buckets include high-volume hiring, hiring and retaining talent, intelligent interviewing and team productivity.

The last area is DE&I enablement, raising the question: How can you turn the data you collect into valuable insights and get better outcomes? The sub-buckets include hiring, retention, sourcing and candidate experience.

The benefits and risks of applying machine learning in recruiting

Using machine learning for recruiting brings uncountable potential benefits but also risks that can negatively affect the hiring process.

Arguably the biggest benefit of tapping into machine learning in your hiring process is the ability to make better decisions quicker. Susanna says, “One of the interesting ways that I've seen it used is to give insights into the interviewer as opposed to the interviewee. Say you interrupted these candidates this amount of times, or maybe there's a pattern here that you might be favoring male applicants, allowing them to speak more than women, things like that.” In these instances, ML models can help companies improve their hiring process by identifying biases like these.

ML models can also pose a risk when it comes to making hiring decisions. Mona explains that the algorithms used to make product recommendations can also be used to make candidate recommendations. “There are often other models that go on under the hood that assume your demographic categories and then use that information to further recommend you products. Suddenly you've got a feedback loop. Something as simple as viewing products on a website and then using a similar algorithm for a much more consequential decision like recommending candidates.” This can have a profound effect on hiring outcomes and poses a risk because the impact and implications of those outputs are unknown.

Greenhouse’s approach on using AI and ML to improve the candidate experience

So what’s Greenhouse’s approach to using AI and ML to improve the candidate experience? Mona says, “We’re still learning as we go. For example, we've largely taken a step back from making direct candidate evaluations and decisions at Greenhouse because we're not yet confident that we can do so without introducing bias that we can't keep track of.” She adds, “If we're collecting EEOC information on candidates, we can use that to ensure we’re not discriminating against candidates based on those categories, but that's not the finality of a human being’s experience. At the moment, we're not confident that we wouldn't disadvantage candidates in a blind spot that we're not aware of.”

When Henry asks the panel about how we can leverage AI and ML technology to create a great candidate experience, the panel has a more positive outlook on the possibilities.

Mona says that using AI and ML to derive intelligence as part of the hiring process could be a great way to elevate the candidate experience. She explains that identifying interviews that are impactful at different points in the hiring process can allow decisions to be made faster, allowing candidates to be selected quicker and removed from the process earlier if they're not a match for the company.

Dave shares that he’d like to see predictive models for how successful the person is going to be, or how happy and fulfilled that person will be at a job. “I would love to see candidates who have not been communicated with in a really long time be communicated with at the exact right moment when we can help them get where they want to be. Those are the types of things that I think are pretty positive.”


If you joined us at Open 2022 and would like more resources, you can access the attendee hub here.