Recently, I've spoken at a number of industry events, including Talent Leaders Connect, and the AGR Graduate Recruitment conference on a subject I’m passionate about: how quickly technology is changing the world around us, and how willing we are to embrace that change.
Can you believe, Apple unveiled its iPhone just over 10 years ago. At the time, there was no 3G network and not a single app in the App Store. And yet today, we’ve come to rely on these devices so heavily that it’s difficult to imagine how we’d function if they suddenly up and disappeared.
This idea really hit home the other morning when I overheard my 6-year-old daughter having a “conversation” with Amazon’s Alexa about whether or not she needed to wear her wellies that day.
Of course, given my chosen profession, I’m incredibly interested in how AI-driven technologies are affecting what we do in our world of HR and recruitment. Whether you’re a Head of Talent or working on the ground as a recruiter, your success depends on your ability to quickly and effectively identify, engage and ultimately hire the best possible candidates for your organisation. As candidate pool volumes and the competition for talent continue to increase, hiring professionals are expected to deliver better results with fewer resources at their disposal.
I’ve spent the last five years looking for solutions to this very problem, and what I’ve learned is that AI and machine learning tools are a key part of the equation. However, leveraging these technologies correctly in a recruitment context can be a challenging thing to do. If we’re going to win the “war for talent,” we need to not only embrace these tools, but also understand how to make them work for us.
Before I dive into how I believe AI will transform recruitment, AI is already making waves in other industries I’ve picked a few examples to highlight how AI is being used successfully already.
An app has been developed that leverages “computer vision” technology, which enables computers to interpret images or videos on their own, to analyse a user-uploaded image of a mole and offer an immediate opinion on whether or not it may become cancerous. Due to its singular purpose and the vast amount of data it has at its disposal, Skin Vision’s diagnoses are extremely accurate in many cases, more accurate than a doctor’s. When you consider the sheer range of conditions a general practitioner addresses, it’s unrealistic to expect them to have expertise in every area.
This app aims to help laypeople understand complex legal documents, uses natural language processing to analyse contracts and generates an executive summary, highlighting the most important points in simple terms.
So how does it work? Machine learning is the process of building algorithms that learn the relationship between inputs and outcomes. Skin Vision’s algorithms are “trained” over time by being shown thousands of different moles and informed as to whether or not they became cancerous. As more and more people use the app and the dataset grows, the degree of accuracy increases accordingly.
So what does this all mean for the future of recruitment? These days, we’re not just up against the high competition and increasing volumes of candidates I mentioned earlier, we also have to overcome our own psychological barriers in order to achieve real success. Human decision-making is inherently biased, which, left unchecked, can result in an inefficient and inconsistent recruitment process that has a negative impact on quality of hire.
When I first got involved in the world of HR and recruitment, I recognised that with the proper combination of video and AI technologies, I could provide a solution that would help hiring professionals make better, fairer, faster and more consistent decisions about talent.
After five years of research and testing, we’ve managed to develop a machine learning platform that helps recruiters utilise their recorded video assessments in a more objective and scalable way than ever before.
Put simply, the platform measures and rates reviewer scoring and identifies potential inconsistencies so recruiters can locate problems in existing processes and make the appropriate adjustments. Additionally, when a reviewer’s score is challenged, the algorithm automatically re-introduces the candidate in question for a second review by other team members to verify the accuracy of the initial assessment.
This exemplifies a systematic way to introduce continuous improvement into an assessment process. Leveraging hard data to drive better hiring decisions reduces the risk of missing out on the “hidden gems” — an incredibly common problem among businesses recruiting high volumes of graduates.
Right now, we’re working on new predictive technologies capable of instantly assessing applicants and identifying high potentials. Based around our video assessments, we start by transcribing every video and analysing each candidate’s language and speech patterns. Algorithms can measure key personality and aptitude indicators like sentiment and lexical complexity to determine not only what it is they’re saying, but also how effectively they communicate their ideas.
Then, we look at the manner in which they communicate. We analyse the actual audio files for things like pitch variation, pacing, pauses and other speaking qualities that could be indicative of performance level. Finally, we interpret the video recording itself, looking at facial expressions, movements, eye tracking and more.
By piecing all of these elements together, we can actually paint a surprisingly accurate picture of each applicant, literally within seconds of their video interview submission.
Of course, measures need to be taken to ensure these predictive algorithms work to curb bias rather than perpetuate it. Predictive models are built upon your current employees and decision-making processes, so if you currently only hire white men, for example, it’s likely a predictive model will only surface other white men as potential high performers.
In this instance, the order of operations is crucial, by developing the technology to address consistency and bias issues first, we can use these better decisions as a solid foundation, then bring in predictive analytics to automate the process and reap the full benefits.
While I don’t think a fully automated recruitment process is on the immediate horizon, I do believe this is the direction in which we’re headed. And as the level of automation in hiring continues to increase in the coming years, it’s the early adopters who are ultimately going to gain the upper hand in the ever-present “war for talent.”
If you’re interested in how automation can be applied to improve the recruitment process, you can download our new guide for recruiters here.