Artificial intelligence and machine learning for recruiters
Some say the industry is doomed — that machines will eventually perform all the tasks that human recruiters carry out today. Others dismiss the technology entirely, saying that AI will never replace the crucial relationship building aspect that recruiting depends on.
So, which is it?
How AI and ML can improve candidate sourcing
Companies are struggling to identify qualified talent in today’s tight labor market. That creates a healthy demand for recruiters to capitalize on.
While professional recruiters can identify talent much faster than a company’s HR generalist or hiring manager, recruiters still need to complete time-consuming tasks and weed through a lot of noise to source candidates effectively.
The most significant impact that AI and ML can make on hiring and recruiting is at the sourcing stage. Recruiters spend a good chunk of their time combing LinkedIn, working the phones, and punching out emails to compile a talent pool to work with. This laborious element of the job accounts for roughly 50 percent of the work recruiters are doing. But how is sourcing success measured? Especially when it’s a fluctuating talent pool between additions and placements. It’s fair to say that sourcing is just the tip of the iceberg regarding recruiting success.
Determining whether a candidate is qualified for the position is one aspect of hiring. For instance, does candidate A have a Master’s degree and more than seven years of project management experience? How much new business has candidate B generated in the past six quarters?
These qualifications need to be met, but they only amount to a fraction of the success equation. The most important considerations should be focused on how well a candidate fits the team dynamic, their ability to grow and handle new challenges, and what makes this a successful placement for the long-term.
How can AI and machine learning help?
Applications can be programmed to sift through resumes, candidate profiles, and applicant data to aggregate a list of the most-qualified candidates. It can also manage blind hiring efforts for building more diverse teams. All the recruiter would have to do is set the search parameters (education, experience, skills, and so on) and let the hype-fast technology do the rest. But any real hiring manager or recruiter knows it’s not that easy.
At its best, ML is being taught to form opinions, and those opinions are optimized to be good at predicting future events based on past events. Talent, by nature, is fluid with unique characteristics or, “features,” as they are called in the ML world, that simply aren’t recorded in the resume or job description.
If you act like a machine, then you will be replaced by a machine. Recruiters add value beyond the resume, and the character fit aspect is an important element that would be difficult for a machine to replace.
Recruiters not only find the right candidate, but sell the candidate on the opportunity – something that could be very hard for a machine to do.
So, how can we give recruiters more time to find candidates that might not have even been looking and sell them on the opportunities? Machines can help free up this time.
The increasing ease in which candidates can find and apply to jobs has undoubtedly increased the noise that recruiters need to deal with. Machines can help and suggest candidates to focus on, but companies should be very cautious to apply automated, “weeding out,” processes.
By using AI and ML to absorb the tedious task of sourcing, a recruiter can free up 50 percent of their time to focus their energy, creativity, and intuition on the recruiting elements that matter most.
Related: Learn more about artificial intelligence and machine learning with our helpful cheat sheet!
AI and ML can’t vet and score candidates
While these emerging technologies offer recruiters an opportunity to improve their sourcing efficiency greatly, a human element is required to vet and score candidates.
A resume and LinkedIn profile might qualify a candidate for an interview, but most of the scoring in the vetting process requires human conversation.
A rising example of a place where employers and recruiters should tread smartly are chatbots. Chatbots, while fun, should be thought of as a user experience to an age-old concept – the FAQ. Machine learning can dynamically expand the FAQ over time and the chat-like experience may feel more natural to surface answers. However, at the end of the day, the bot isn’t designed to have a two-way conversation – it’s designed to listen to the candidate, determine what they asked, and align that with a predefined answer.
This is great for many scenarios and can save the candidate and recruiter a lot of time, but at some point, the conversation is going to move past this to one that requires decisions to be made and trust to be established. That is where the recruiter comes in. This is where having more time for those conversations is critical.
Granted, some ML applications can go beyond keyword hunting and are capable of contextual research. With the proper permissions, the software could theoretically conduct a criminal background check and confirm education and employment. But how would it accurately consider the vital social elements of recruiting?
Could it determine the type of culture a candidate works best in? Or if a company’s current structure supports the candidate’s career goals? Would it know about recent mergers and acquisitions that will affect a candidate’s interest?
Probably not. To execute the sub-surface elements of the recruiting process, a skilled, experienced human recruiter needs to be behind the wheel.
TIP: Check out the best machine learning software for your recruitment needs!
The future of AI and ML
With the damage of a bad hire being so pronounced for organizations, AI and ML applications would have to demonstrate a robust recruiting record before being dependable at executing a complete search.
In the recruiting world, reputation is everything. It’s difficult to cultivate a positive reputation for recruiting excellence, and a couple of poor placements can ruin it all.
Presently, AI is a very exciting field with a lot of potential. That potential should be focused on empowering candidates and recruiters to make better decisions and connect more easily.
Over time, this field will advance, but until a recruiting AI can pass the hiring equivalent of a Turing Test and demonstrate judgement and insight, hiring is probably best left to the humans.
Want to read more about the Turing Test? Check out our comprehensive history of AI!
Aaron leads Crelate’s customer-centric approach to product design. As a self-taught developer, experienced consultant, and visual arts major, Aaron brings a blend of design, technical skills, and customer empathy to the products he creates.
Crelate is Aaron’s fourth successful product. The first was sold to Microsoft in 2001, where he led the Application + UX development teams for Microsoft Dynamics CRM 1.x to 3.0. After, Aaron and his co-founders started a technical consultancy, which grew quickly to almost 150 strong. During this time, Aaron experienced first-hand the challenges of recruiting and building a talent-based business. This business was sold to Avanade in 2010.