The application of artificial intelligence to traditional business processes has led to a massive shake-up of the employment market. Shareholders are told that AI can perform tasks quicker and more efficiently than humans, so it’s no surprise when workforces are reduced and payrolls cut in order for those companies to remain competitive.
But this isn’t the only area of employment where AI is having a big impact. Yes, many repetitive and basic tasks previously performed by people are being automated and run by AI – but AI is also a job creation machine. Take for instance the software industry. Many junior-level coding and testing roles are disappearing thanks to AI – but new jobs are being created too: machine learning specialists, hybrid engineering roles, and in particular, data scientists.
Yet because these roles are relatively new, there aren’t enough people with the requisite skills and qualifications to fill them. Ironically, while lay-offs are happening in one part of the industry, another part is desperate to find the right employees for this next-gen workforce. HR departments are caught in the middle of this problem, expected on the one hand to help people out of the door, while on the other to deal with an emerging AI talent crisis, where they’re forced to participate in a frantic – and increasingly expensive – scramble for suitable candidates.
Outdated hiring practices fueling the talent gap
But is the real problem a disconnect at the heart of how HR works? Are outmoded hiring practices leading to decisions being made based on only the most simplistic of employment data? And is the talent crisis itself an artificial construct that’s merely the result of an inability to measure and assess the people that companies already employ?
In our experience, most enterprise HR departments are still using systems designed for a pre-AI world, predominantly based on screening CVs for relevant credentials and prior experience. This approach made sense when roles were relatively stable and learned skills had a reasonable shelf-life. But in a world where companies are hiring for an uncertain future, with job specs changing from month to month, this approach is no longer fit for purpose.
The mass availability of generative AI apps such as ChatGPT has also created a volume problem too. AI-generated applications have flooded recruiting pipelines, pushing HR toward the bluntest of filters: degree requirements, title matches, years of experience etc. Yet these are exactly the type of criteria that fail to identify the best candidates when the role being hired for requires skills and aptitudes that exist at the cutting edge of the present rather than rooted in the past.
For example, when a company we had discussions with tried to fill an AI workflow management role, the HR team defaulted to requiring a degree and three years of an identical job title, even though that role barely existed 18 months ago. Despite internal candidates showing strong learning agility, their system filtered them out. That experience epitomises how HR still operates: measuring the past rather than understanding future potential.
The artificial divide between redundancy and recruitment
Companies are laying off people whose specific tasks are being automated, while simultaneously struggling to fill roles in machine learning operations, data governance and AI workflow management. These are not fundamentally different populations of people, yet organisations treat them as completely separate workforce pools, one a liability, the other a valuable commodity. This is at the crux of the talent crisis that many companies now face – an artificial gap between these two pools due to a fundamental failure of measurement.
If a company cannot reliably assess whether its existing workforce has the cognitive agility, learning velocity and underlying reasoning capability to make a transition, it will always default to the external market — not because it wants to, but because the external CV is at least a familiar signal, however imperfect. But what if HR was able to work smarter, and instead of writing off capable employees, it could identify which of them, with the necessary retraining, might be redeployed over expensive external hires?
That’s not to say that retraining in itself is a silver bullet. In fact, many companies are more than willing to try this approach. The big issue is selecting the right internal candidates to go through this process. Today, most organisations lack any reliable mechanism for identifying who within their workforce has the underlying capabilities to make the leap to a new role versus who will struggle and stall. Without that, retraining programmes become expensive and slow, and HR leaders reasonably conclude that external hiring still produces better results.
Why retraining often fails
For example, we had a case where we were retraining accountants into data-adjacent roles. While most of the candidates struggled, two excelled. Subsequently reviewing their profiles, we realised that those two candidates had high learning velocity and the motivation to pivot to a new role. But without a way to measure and identify such candidates upfront, HR defaulted to external hiring. That experience clearly outlines why retraining often fails in practice: the process is entirely hit and miss because organisations are not able to accurately identify who to bet on.
Yet as the supply of external candidates for AI-related jobs continues to be extremely pinched, internal redeployment remains an important way out of this problem. The organisations that get this right will be the ones that invest in measuring potential directly eg. cognitive ability, learning agility, motivation to grow into new domains etc, rather than inferring it from qualifications or past job titles. Once HR can accurately assess these qualities, the build-versus-buy decision becomes rational rather than a gamble. They can retrain and redeploy internal candidates whose potential they can now see, and hire externally only when there’s genuinely no other option.
A failure to recognise the potential flexibility of internal candidates has helped create an artificial talent crisis, with lay-offs hurriedly pushed through without a proper appreciation of workforces’ ability to adapt and pivot. The key to escaping this relentless cycle of fire-and-hire is through seeing each employee as a fully-rounded person and not just a set of outdated credentials.
Kevin Chan is CEO of Epitome Global, a workforce optimisation company that uses whole-person intelligence — skills, motivations, values, preferences – to help enterprises and governments understand the full potential of their employees.
