AI has been everywhere since ChatGPT went mainstream last November. It has dominated the conversation in the recruiting space, with talent teams trying to figure out how it can be successfully integrated into their talent strategies.
Our recent Untold Stories webinar took a deep dive into the ways that AI can elevate the recruitment process.
Anthea Hartzenberg, our Head of Engagement, spoke with Jeet Mukerji, the CEO and co-founder of an exciting new PeopleTech stealth startup. They discussed how AI-integrated applicant tracking systems can be used for skills inferencing, helping hiring teams unlock otherwise unseen, high-quality tech candidates.
A demystified look at AI and its applications for talent teams
We’ve only just begun to scratch the surface of AI and how it can revolutionise the recruitment industry. Jeet is bullish about the partnership between AI and recruitment, outlining how talent teams are already using it:
- To set the direction of the recruitment process through the combination of key parameters, data sets and human judgement. There’s no silver bullet coming to replace the process or HR as a whole; AI is a collaborator alongside existing internal processes.
- As an efficiency enhancer. AI can already eliminate routine tasks, including outlining job descriptions, or even assisting with writing general emails.
- For fast talent sourcing, albeit not through total automation. AI is not powerful enough to make hiring decisions, but it can help HR teams do so faster. It can sift through 1000 applicants faster than a human is able to, but nuance inputs from the latter remains highly important.
A breakdown of AI-powered skills inferencing and why it matters
Much of the discussion centred around skills inferencing. Jeet explains:
“Skills inferencing is when you can infer skills from a collection of other contexts around it.”
The AI essentially emulates what we humans can do in our heads already, namely seeing patterns from earlier data or other contexts, to make rational deductions.
A practical example
A successful hire depends on matching the right skills and experience to the vacancy. Rather than just trawling what’s written on a CV, AI-powered skills inferencing analyses the data in relation to other data sets, such as the seniority level of past roles or if the candidate’s current company culture matches that of the hiring company.
It also takes into account what’s currently happening within the hiring company, such as the growth rate and what its goals are, to provide an assessment of whether the candidate is a good match. For example, AI can trawl data to see whether a product manager is or has worked for a company that has displayed tangible growth or not (through public performance records, investment numbers, etc).
Driving better sourcing goals with your ATS configuration
Applicant tracking software (ATS) is one of many efficiency-driven recruiting tools in the industry today. Jeet suggested that these are traditionally tailored towards a job-specific view. HR teams are purely looking at candidates’ skills and experiences in relation to the needs of the position. Despite best intentions, good candidates are still left behind.
AI can and will change that. A major limitation — as is the case with ChatGPT in its current form — is that its input data can quickly become out of date. ChatGPT itself currently only works from input up until the end of 2021. In a fast-changing, highly globalised world, this can significantly alter its accuracy.
“At the end of the day, your AI is only as good as the data you put into it.”
Jeet believes that AI integration into ATS requires highly personalised data sets in order for it not to default to generic responses. Inputs need to be relevant to the needs of the role and the business, or the results may be unreliable.
Bias and neutrality
At its optimal performance, ATS will meet AI in a bias-free way. Anthea shared that Amazon’s AI engine from 2018 — designed to surface top talent regardless of gender — was highly skewed to male candidates. It rated candidates through inherently biased input data over a decade when most resumes came from men.
Another example related to hiring managers using tech to whittle down candidates based on the reputations of their universities, exacerbating the problem of hiring for privilege and not necessarily skill.
“If you want to be really unbiased, then it’s worth looking at how we might layer AI on top of the ATS through another tool that may be more candidate-specific.”
According to Jeet, this adds another dimension into the mix, but it doesn’t change the fact that getting the most out of AI-powered ATS requires constant tweaking until the business finds the right balance for its needs.
Broadening talent pools with AI-powered ATS
Broadening talent pools with AI-powered ATS requires taking a step back and considering how the AI can integrate within the context of existing processes. Consider how you can make these better, as well as the following:
-
Hiring is a team sport
AI-assisted hiring should always be collaborative. In order to broaden the tech talent pools available for future hires, recruiters and developer teams need to work closely together. They should always be on the same page about how data is used, when it is updated, and what the inputs are.
-
Readily available internal skills
AI can broaden the available talent pool without even having to make a hire. It can be used to uncover previously hidden skills already present in the business.
“We can surface hidden talent internally rather than just looking externally.”
Jeet suggested that this gap analysis has the two-for-one benefit of altering the hiring strategy to find what the business actually lacks while giving employees the chance to grow based on existing skills they aren’t actively using in their current roles.
-
Bank talent pools for later
AI allows recruiters to group skills – such as most desired, desired and least desired –which Jeet sees as integral to expanding access to talent pools. Once a role is filled, especially in a high-volume recruiting environment, the top percentage of candidates can be used for HR teams to identify new prospects for future hires.
“So your time to hire essentially goes to zero, because the talent has already been identified and you’ve already built up a relationship. Now you just need to wait for the job to open up and ask them to apply again.”
Jeet confirms it’ll make hiring a proactive process rather than a reactive one.
Elevating recruitment using AI still requires human input
AI is not a “ready-made” solution that will take away the need for recruiters. It actually requires a lot of input and interaction to support and elevate existing internal processes.
Read more
- The Best Practices for Hiring Developers. This chapter from our Ultimate Developer Hiring Guide unpacks best practices to set yourself up for success throughout the hiring process, and ensure a good experience for both candidates and your team.
- How Recruiting Tools Can Streamline Your Tech Hiring Process
- Four Considerations for Procuring Equipment in Time for a Developer’s Start Date
