AI, or artificial intelligence, is revolutionizing everything from healthcare to retail. The world of recruitment is no different. As human resource managers begin to deploy AI for live job interviews and to assess other candidate questions and answers, this tool can dramatically decrease time to hire, as well as help companies improve their candidate matching abilities. Here’s how AI can help recruiters improve the hiring process by using automated interviewing and other tools.
What We’ll Cover
- What is AI recruiting?
- What role does AI play in HR?
- Why recruiters are switching to AI recruiting?
- Human vs. robot: why AI is winning
- Cost savings to HR teams
- How AI interviews work
- What jobs can you use AI interviews for?
- What questions can AI grade?
- Can candidates trick AI into hiring them?
- How to set up an AI hiring process?
- AI works best at scale
- Ethical considerations of using AI
- What’s next? Candidates deploying their own AI to pass a job interview?
- Additional Resources
What is AI recruiting?
AI recruiting is the practice of integrating artificial intelligence, such as machine learning or predictive algorithms, to the time-consuming or resource-intensive parts of human resources recruiting.
AI can be used by recruiters to streamline some of the repetitive tasks that make it difficult for small recruiting teams to quickly identify and contact their top candidates. Today, recruitment teams apply AI for three primary scenarios.
An AI tool can search for resumes and profiles on the internet or in an internal database using keywords that fit an open position, surfacing all the possible best fits. It can also provide real-time messaging and automated updates to candidates as they move through the hiring funnel. Some AI tools can interact with a candidate to determine what specific role will be the best fit and show them how to apply through the job site.
An AI tool can screen resumes to elevate the candidate who looks best on paper, based on keywords and other qualities. Other AI tools use skill tests results to predict performance. By assessing a real-world simulation, these AI tools have the ability to overcome some of the human bias that sometimes disqualifies candidates from making it to the next round. Predictive performance based on skill testing tends to be a better way to match candidates with open positions, given that a skill test indicates current knowledge and ability, versus the historic nature of resumes. Either way, AI tools streamline the process and help screen candidates in, not out.
AI tools are used in interviews in two key scenarios. In the first instance, companies like Apple, Google, and Facebook have begun to use AI to assess pre-recorded video interviews, using voice and facial expression analysis to assess personality traits. While this can help cut down on the time and attention required of recruiters to review each candidate’s recorded response, there are some big red flags with using facial recognition to select candidates – more on that in a minute.
The second way to use an AI tool with a pre-recorded video interview is to analyze the content of the answer. For example, a skills test may ask a candidate to write a sample blog post on a topic relevant to the company’s industry. Then, in a pre-recorded video interview, the tool asks why the candidate selected the topic they chose to write about. The video portion can be used to give more context to the candidate’s skills test and help a candidate stand out beyond their initial response.
Want to learn more about skill testing?
Hire the very best by focusing on who can do the job, not who looks good on paper. Speak to one of our experts to learn how to embed skill testing in your hiring funnel.
These uses of AI in recruiting are just the tip of the iceberg. AI, when utilized correctly, has the potential to help companies hire on true merit, rather than a list of qualifications on a resume or connections a candidate may have in their personal network.
The most promising scenario for many recruiters is the ability to use AI in their hiring process. Integrating an AI tool can reduce metrics such as time-to-fill, employee turnover, and candidate attrition. It can use performance predictors to help the best candidates stand out on their own merit. This guide breaks down some of the benefits, as well as challenges, of integrating AI into the candidate interview experience.
What role does AI play in HR?
AI is a powerful tool for helping HR team find great candidates, and, in the process, improves some of their KPIs. Beyond this tool’s ability to automate low-level tasks, AI recruitment can help hiring teams improve their quality of hire, save time and money, and minimize some of the hidden human bias that prevents minority candidates from getting a fair evaluation.
By some estimates, the total cost of hiring one new employee could be as high as $5,000 or more, depending on the industry. Candidate interviewing is one of the more expensive parts of the overall recruitment process. Quick math shows that on average, companies likely spend $100 per candidate per interview; pre-screening alone costs an average of $40 or more per candidate. That’s without factoring the costs of interviewing a remote candidate. While video interviewing can help recruiters save on travel expenses, it still takes a lot of time and effort to watch each candidate’s pre-recorded response.
AI is being used to help reduce some of the costs of automated interviewing by screening candidates. Existing tools can assess the candidate’s interview by understanding:
- Analysis of facial expressions, body language, and gestures
- Voice and text sentiment analysis
- Actual ability to do the job
- Cultural fit and personality
AI can help recruiters expand their candidate pool beyond the initial application submission step. It’s a great way to screen candidates in, rather than out – and as a result, more and more recruiters are switching to AI recruiting.
Why recruiters are switching to AI recruiting?
The competition for talent is only increasing as the job market continues to favor applicants. Nationwide unemployment is at a low 4%. Companies must make extra effort to stand out to top candidates, as well as expand their hiring funnel to reach more potential new hires.
According to industry analysts, 52% of talent acquisition leaders say “the hardest part of recruitment is identifying the right candidates from a large applicant pool.” Not only is it difficult for recruiters to screen effectively, but 56% of those surveyed reported that they expect their hiring volume to increase. Yet, 66% of those recruiting teams will either stay the same size or shrink. In short, companies are hiring more people with fewer human resources available to make sure the candidates are the right fit.
AI recruiting is the glue that makes all this movement possible. In addition to sourcing tools and recruitment management systems that make it possible for agile HR teams to keep ahead of hiring demands, AI is a critical part of processing the influx of candidates in the hiring funnel. Not only does it speed up the interview process, but AI can actually lead to better hires. When trained correctly, an AI tool can often predict on-the-job performance than some human resource managers.
Human vs. robot: why AI is winning
Bias in recruiting exists for many reasons. Primarily, it’s a question of efficiency: HR teams may receive hundreds of resumes for a single open position. It’s difficult to give every candidate’s application careful consideration – and as a result, many recruiters advance candidates whose backgrounds seem similar to their own or more familiar. Those candidates with different backgrounds or work experience never stand a chance.
AI can make a difference in how candidates are processed through the hiring funnel. One such algorithm allows HR teams to evaluate 10,000 candidates in the same time it takes for an individual recruiter to assess one. An AI tool is able to screen candidates based on what they can actually do, rather than what their resume says they can do, evaluating their video interview more efficiently and more effectively than a human recruiter.
And finally, not all interviewers are as talented at providing a great candidate experience as they might believe. Interviewing is hard, and awkward interviews are more common than one might think. The reason for this? Most interviewers allow biases, not facts, to lead the discussion.
AI is a powerful way to remove bias from the equation. The results speak for themselves. Some AI tools have reported a 20-100% increase in gender, ethnic, and socioeconomic diversity of hires. Of course, there are some examples of an AI recruiting tool failing to prevent bias. When this happens, the flaw isn’t in the technology – it’s the data being used to feed the algorithms. We’ll revisit this issue later on.
Cost savings to HR teams
HR teams can clearly save a lot of time and resources by integrating AI into their interview process. There are also significant hidden costs that teams can mitigate by adding automated interviewing software.
Early stats show that recruiters that use AI see a 30% reduction in cost-per-hire, as well as increased rates and shorter time to fill. The ROI of using AI in hiring includes some of the long-term variables which predict a candidate’s success. Consulting firm PwC points to “not only the organizational ROI in hiring, training, and remuneration, but also relating to the individual’s ability to learn on the job, develop new skills, and his or her social contributions.”
The costs of hiring the wrong candidate are steep. A 2016 CareerBuilder study found that 75% of employers participating in the survey said they hired the wrong person, costing companies an average of $17,000. The cost of making the wrong hire can be up to 2.5x salary. Where candidates routinely lie on their resume or human bias interferes during the hiring process, an AI tool can help cut both the hidden costs and opportunity costs of hiring the wrong new employee.
How AI interviews work
There are a variety of AI interview tools that assess a candidate’s pre-recorded video questions and answers. Essentially, these tools use algorithms to understand data points from the interview, including:
- Analysis of facial expressions, body language, and gestures
- Voice and text sentiment analysis
- Actual ability to do the job
- Cultural fit and personality
One such tool examines 25,000 different data points per video – everything from tone of voice and cadence to facial expressions and words. Unfortunately, this type of machine learning can cause major problems for recruiters. Instead of assessing ability, a tool that seeks to hire people based on facial expression reduces diversity by promoting “look alikes” of existing employees. Modeling based on top performers is a flawed strategy that stifles innovation and growth.
AI tools work better when their purpose is to look at content and behavior from a skills test. The data being fed to the AI tool should inform the algorithm on performance on specific tasks, rather than how a candidate looks and sounds. Using results that simulate the job can increase diversity by giving all candidates an equal chance to showcase skill – a win-win for the candidates and the company.
Deep learning and predictive analytics help the AI tool understand the context, as well as multi-part or changed answers. As more and more candidates participate in the skills test or interview process, the AI tool can recalibrate. Every new piece of information that it gets makes the algorithm smarter, meaning the platform can then give candidates a specific ranking, rather than a pass/fail score, to help the recruiter decide who makes it to the next phase.
Other automated interview tools are a little more straightforward. These tools are limited to guiding the interviewee through the process, rather than assessing the result. In this scenario, an applicant is automatically invited to participate in an interview. Upon accepting, they are provided a link to connect to the system and given instruction. After a microphone/camera test, the questions begin with parameters regarding timing or other associated limits already in place.
The full recorded interview is submitted to the employer for evaluation, and most of the time, the AI tool does have the capability to rank applicants based on their responses.
Benefits of AI interviews
There are many benefits to deploying automated interviewing. Here are just a handful.
Remote work is on the rise, and, accordingly, so is remote hiring. It’s becoming increasingly common for a workforce to be spread across the globe. Automated interviews can help recruiters keep pace with this trend. When a hiring team is limited to in-person interviews, their hiring pool becomes even more shallow. Automated interviews let recruiters easily find out whether, and how, each candidate can do the job, regardless of location and without wasting time.
Hire at scale
Recruitment management platforms and powerful sourcing tools are channeling more and more candidates into the hiring funnel. It’s great the recruitment teams have more resumes than ever to evaluate; however, the screening process becomes more and more difficult the more applicants there are. An automated interview removes the need for a costly phone screening step and streamlines the process for a candidate to reach the next phase.
Where an automated interview most outperforms the traditional interview experience lies in the aptitude testing inherent in the format. Automated interviewing is a great way to see how a candidate performs on-the-job tasks. AI tools can assess a mock sales pitch, a customer service call, or a presentation using Powerpoint slides. Automated interviews help understand how candidates do the tasks that matter to you most for the job you’re trying to fill.
Offer a great candidate experience
Automated responses not only save time for the recruiter, but also for the interviewee. Most candidates don’t want to waste time setting up interviews, traveling to the employer, and participating in endless phone screens. The improved transparency and quick communication built-in to automated interviewing deliver a vastly improved hiring experience.
Pitfalls of AI interviews
There are few pitfalls to using AI interview when the technology is utilized correctly. However, there can be some drawbacks to using automated interviewing when the tool isn’t being fed the right data.
The biggest issue that arises is when some AI tools are used to interview candidates is in multiplying or replicating human bias. The notorious example of Amazon scrapping their AI recruiting tool after it was discovered to have been replicating the biases found in the human recruiting processes is just one example. The flaw in this tool wasn’t the technology; it was the data being used to feed the algorithm.
Tools that use facial recognition or voice analysis are the biggest culprits in preventing diversity hiring. It’s crucial to give AI tools information outside of the resume – interview questions and answers, for example – to get a different result than if a human recruiter ranked candidates manually.
What jobs can you use AI interviews for?
AI interviews can be used for virtually any industry, but some positions are harder to assess than others. According to one expert, professional industries with fewer candidates, such as IT, nursing, and senior-level positions are more complicated to hire for – and therefore may not be the best places to use AI interviews.
Key to using automated interviewing tools is to understand what success looks like in an open position. For example, if you’re hiring a call center rep, the AI tool should be assessing voice cues and qualities like patience, rather than facial expressions. For programmers, an AI tool can prompt coding challenges in a certain language or skill set. Learn what competencies are important for on-the-job performance before codifying the assessment tool to give you the best candidate.
What questions can AI grade?
AI has the versatility to grade questions and answers for many different types of roles. From coding challenges to manipulating cells in Google Sheets, the automated interviewing tool can handle a variety of skill tests and cognitive assessments.
Here are just a few examples of questions companies have put AI in charge of evaluating.
AI can evaluate coding challenges, Powerpoint presentations, and more:
Can candidates trick AI into hiring them?
It’s difficult to trick an AI tool, especially when used in a skills test, as the purpose of AI ranking is mostly to screen candidates in, rather than out. However, some machine learning tools have been set up exclusively to prevent cheating on one-way interviews, for example. One automated interview tool detects if the candidate is regularly looking away from the screen (and possibly relying on cue cards) to answer the question. Or, the tool can pick up if there’s another voice on the recording coaching the candidate to answer questions.
Regardless of the tactics that some candidates may try, inevitably the AI tool is part of a larger recruitment process. If a candidate “cheats” or tries to outsmart an AI tool, a recruiter should be able to triangulate a skills assessment with other facets of their application. It’s all part of creating a formula that helps the right candidate stand out from the crowd.
How to set up an AI hiring process
There are three main ways to install AI in your hiring process. The first is as a resume screening tool in your existing database. This tool can learn over time which candidates continue to be hired and stay on with the company, using metrics like performance, tenure, and turnover to learn what skills and qualities to look for in future applicants.
The second way is as a communications tool. Some smaller hiring teams use AI to automate the setting up of interviews, reject applicants, or alert candidates when their application makes it to the next step of the process. The main benefit is in improving the candidate experience by adding transparency. And, when 58% of job seekers report a negative impression of a company if didn’t hear back after submitting an application, transparency can make a huge difference in a company’s recruitment efforts.
Finally, and most impactfully, AI tools can rank skill tests and video interview responses. The most comprehensive AI tools manage the candidate experience from start to finish, automatically grading candidates based on how they perform, evaluating at scale, and keeping in touch with applicants throughout the process. These platforms come with comprehensive integrations that make it easy to add AI with your existing HR tools.
AI works best at scale
Does AI work for all HR teams?
Yes and no. While the technology is the same no matter what size HR team, a smaller organization may benefit more from having human recruiters manage the interview process. AI tools need a statistically significant amount of data to avoid the bias trap; the tools become more powerful the more input they can learn from. In addition, small organizations will be better served by using their size as a strength. The hiring experience must meet the employee experience. Culture fit feels slightly more important when people are working at a smaller company. As a result, smaller HR teams might benefit from building personal relationships with each and every candidate who comes in to interview.
Ethical considerations of using AI
Is it ethical to use AI to make hiring decisions? In some ways, automated interviewing is a vast improvement over our existing hiring practices. When the data being fed into the AI algorithm is based on performance, the ability to remove human biases and match candidates on merit is huge. AI tools that get the proper data sets have the potential to change the hiring market for the better.
When recruiters use AI to screen candidates in, rather than out, hiring becomes more democratic and merit-based, rather than about who an applicant knows. Companies who switched to one AI tool saw a 62% increase in female candidates. Nepotism and selective hiring practices get eliminated when an AI tool is correctly utilized.
What’s next? Candidates deploying their own AI to pass a job interview?
Will hiring soon be delegated exclusively to robot representatives? Probably not in the near future. But, according to one AI trendwatcher, it may not be ridiculous a proposition.
In some ways, recruitment tools have already ushered in the possibility of constructing a candidate-side AI tool. A recent article by TheVerge outlines how prospective job applicants have already been reduced to a “series of data points.” Between a candidate’s social media profiles, LinkedIn, and internet history, an employer has a rough outline of a person’s ability before the interview process. It’s not far off to suggest that a candidate may attempt to utilize AI to construct their own digital representation, consolidating their internet presence in a way that best represents them to a recruitment algorithm.
For more reading, check out some of these great resources.