AI Will Dominate Recruiting – So Prepare For Major Changes In These Areas

It’s quite common these days for the CEO’s from Amazon, Google, MS, Facebook and Apple to expound on how artificial intelligence and machine learning will dominate their businesses over the next few years. Even Vladimir Putin stated, “The country that leads in artificial intelligence will lead the world.” It’s also important to realize that in addition to contributing to the most visible product areas, like digital assistants and driverless cars, “Machine learning and AI are a horizontal enabling layer” says, Jeff Bezos of Amazon, meaning that AI will impact and improve every major function and its processes and decisions. Recruiting leaders shouldn’t be surprised that I predict that “machine learning will soon begin to dominate every major aspect of recruiting.” Just as previous technologies like ATS’s and CRM’s have already transformed recruiting. It’s important for recruiters to be aware that there is an upcoming wave of mostly vendor developed recruiting applications that assist in producing extraordinary hiring results because they include machine learning capabilities.

The goal of this article is to highlight the upcoming AI/ML and technology changes that are likely to occur in each of the major areas of recruiting.

 

The Top 15 Recruiting Areas That Will Be Most Impacted By AI And Machine Learning

The areas of skills-based recruiting and job/candidate matching that will be impacted are below. Note that they are listed so that the initial items in the recruiting process appear first.

Recruiting areas related to finding and attracting prospects

• Advertising placement and content – Machine learning will continually improve your placement process for branding materials and job postings rather than relying on costly trial and error approach to advertising. This is critical because accurate placement is essential if you expect to get the right kind and number of applicants. Systems will continually learn by analyzing visitor cookies and response rates so that you place your highly targeted materials in front of the right people at the right time. Also, machine learning technology can help you continually refine your content so that it gets the highest response from your recruiting targets.
• Your own website and social media – continually improve by firms using machine learning on their web and social media pages to better attract and continually engage your target audience. Software bolstered with machine learning will also be able to monitor and make you aware of both positive and negative comments that others make about your firm and jobs on the Internet and social media.
• Finding individual prospects – during sourcing will become much more automated and accurate when augmented with machine learning capabilities. Automated sourcing programs will be able to find many more and better matches, based on the continually updated target profile that you develop as a result of feedback. There are already vendor packages that allow you to identify currently employed individuals (e., passives) that are likely to quit soon and prospects that are likely to be diverse.
• Enhancing prospect profiles – can make the existing candidate profiles found on sites (like LinkedIn) more complete by supplementing them with additional information that a machine learning program will find on the Internet. Machine learning driven programs can sort through a prospects search histories, cookies and social media sharing. The additional information on a prospects interest, capabilities and behaviors might indicate that a candidate can do things that they haven’t done in the past. Once they apply, chatbots can contact an applicant directly to clarify unclear elements in their resume or profile.
• Improving job descriptions and postings – Recent research data has revealed that job descriptions and job postings can be dramatically improved so that the content better attracts your target audience. So, rewriting them can reduce terms that create a bias. Software can now help you reduce those biases and add content that draws initial attention and that attracts more qualified applicants.
• Responding to questions – from potential or actual applicants is immensely time-consuming for recruiters. So many firms are already utilizing chatbot’s to answer questions quickly 24/7. The U.S. Army, for example, has been using its Sgt. Star chatbot for over ten years to answer its extremely high volume of questions. Chatbots can also periodically update a candidate status, once again saving recruiters time.
• Personalize selling – Machine learning uses big data to identify the attraction factors and the elements of the firm’s employee value proposition that best engage certain personas (e., types of individuals). Rather than a “one-size-fits-all” approach, this allows you to make your attraction, marketing messages and personal communications more effective because they are highly personalized to the individual.

Recruiting areas after candidates apply

• Resume sorting – with machine learning software uses the resumes of successful hires at your firm to find patterns and then it can use these past success patterns as a basis for predicting which resumes and candidates are most likely also to be successful when hired. If programmed correctly, resume sorting software can also help to eliminate a great deal of unconscious bias in resume screening and candidate slate selection. Machine learning assisted search programs can also help you find hidden or lost talent within your ATS database.
• Matching people and jobs – Using matching programs supplemented by machine learning can help a firm determine if there are any, less obvious, jobs that an applicant would also qualify. Matching people with jobs will also be improved by looking not just at an applicant’s past job titles and degrees, but also at their skills and capabilities.
• Interview scheduling – is time-consuming and dramatically reduces your speed of hiring. Fortunately, there is existing software that allows a candidate to self-schedule their own interviews depending on their availability.
• Interviews – can be time-consuming, so it makes sense to automate the initial ones with a chatbot that provides personalized questions based on your job profile. Also, there already exists technology that allows the use of neuroscience tools like voice and facial recognition to assess aspects of video recorded interviews that no humans could detect. There are even voice modulation programs that can help you obscure the voice of telephone interviewees so that it’s harder to identify their gender and national origin.
• Supplemental candidate assessment – in addition to traditional interviews. Natural language processing can check language skills and online technical tests and challenges can help to assess the skills of applicants. There are automated programs that can more consistently determine cultural fit. Eventually, virtual reality simulations will be able to supplement interviews by giving candidates actual problems from the job to solve.
• Offer acceptance – based on the candidate’s persona and profile. Recruiters can put together offers that are more likely to be accepted while at the same time treating all genders equally when it comes to compensation.
• Learning from hiring failures – By definition, machine learning processes continually identify mistakes and errors. Recruiting will have an ongoing failure analysis process that continually and automatically finds hiring and bias errors and their root causes, allowing recruiting processes to improve at a much faster rate.
• Other technologies – in addition to AI/ML technologies. Block Chain may eventually make checking educational and employment credentials easier and more accurate. Skype and video technologies already make it much easier to interview remote candidates without requiring them to travel. Machine learning will make predictive analytics in the area of projecting the future trajectory of finalists (in the areas of performance, retention and promotions) much more accurate.

Final Thoughts

Although most firms don’t track it, the average failure rate of new-hires at all job levels hovers around 50%. For example, Leadership IQ found that when “they tracked 20,000 new hires, 46% of them failed within 18 months”. Former Harvard Professor and author Michael Watkins reveals that “58% of the highest-priority hires, new executives hired from the outside, failing in their new position within 18 months”. Part of this broad failure results from overworked recruiters, normal human errors and unconscious biases throughout the recruiting process. Fortunately, the machine learning technologies highlighted above will soon minimize those problems through automation and continuous improvement. The results will be hiring faster, lower cost and more importantly hires that perform better on the job (i.e., quality of hire), that are more diverse and with fewer hiring failures. Recruiters should also take note that as more recruiting transactions are automated, it will allow current recruiters to “raise the bar” and to move into the more strategic Talent Advisor role.

Finally, recruiters should also be aware that they will soon be recruiting many more individuals into machine learning roles. The share of jobs requiring AI skills has grown 4.5 times since 2013 (Source: Stanford).

Want to see how machine learning can help you find better technical matches for your open roles?  Check out CodeFights Recruiter or attend an upcoming webinar.

 

If this article stimulated your thinking and provided you with an accurate picture of the future of technology in recruiting, please take a minute to follow or connect with Dr. Sullivan on LinkedIn.

© Dr. John Sullivan 5/2/18 for Codefights