Skip to content


Submit or approve timesheets here by selecting one of the options.

AI in Recruitment - an interview with David Mason

The problem today in reacting to developments, especially in technology, is that by the time you've reacted, it has already changed again. That's why David Mason, Director of Talent Acquisition and Management at ConTal - Connecting Talent, stresses the importance of a proactive approach.

AI in RecruitmentClearly following his own advice, David has signed himself up to try out Amy Ingram, an AI virtual assistant, currently in Beta testing. Amy's capabilities currently include setting up meetings based on a user's digital calendar, but that's just the beginning. And though many envision a world where AI will lead to unemployment, singularity and the rise of the machines, David claims that this is not the case.

"The more forward thinking believe it’s more likely to change the nature of work. It’s likely to affect around 30 per cent of our jobs, particularly in terms of knowledge work," he says.

It is also going to affect recruitment in a number of ways, one being talent resourcing, and social media has already allowed recruiters to identify the right type of people to make contact with.

What AI can do is look at patterns in CVs, like the frequency a person moves roles and therefore how and when they are likely to be open to approaches. "AI can do that for us, ranking opportunities and individuals that may fit the criteria. On a mathematical level, AI is better able to predict outcomes from complex sets of data," he continues.

The second point at which it becomes useful is within selection. What AI starts to do is look at a whole range of external factors such as the individual's background, their CV, what they've done before, how they've scored on psychometric tests and then relate those to information around cultural fit. This helps recruiters to get a much more exact science about a person and the likelihood of them being successful in the role.

One of the companies that David’s working with is focusing on the idea of somebody else managing other people's social media profiles for self promotion. "There are Silicon Valley companies that will look at your social media profile and basically, if you don't rank a certain score, they won't look at you," he says. "You've got a whole mass of information coming at you now so there are new tools on the marketplace that allow us to look at the talent in a particular area."

Currently, David is looking at cybersecurity experts in banking and has access to a tool that will tell him how many there are within a 20-mile radius, along with what their skillsets are, their average wages, and the length of time it would take other companies to recruit them. The technology, David says, has taken recruitment down to a 30-day cycle time.

"It's a much more thorough process - we know that we are going through potentially embedded talent but it also means that we can tell the hiring management and executives there is no point in working further in this particular area, as there are 200 people available for this job. ‘We've sifted through 170 in the last 12 months, so your options are to: pay more, go to a different area or move the work.’ The discussion is based on data and evidence." he says.

The use of AI right now is more about understanding than replacing what a human can do. "It's an adjunct to that."

Currently recruiters can go to a job board, Facebook, LinkedIn, or network through individuals. Adding AI is just being more efficient. Most modern recruitment departments won't even touch a fraction of the market without such tools, tells David.

If a recruiter is using advertising to score somebody then around 20 per cent of the market will be active in looking for a job. If they go to a passive arena such as LinkedIn they might get another 20 per cent but the big question is how to get onto the next 20 per cent.

"We are much more efficient about using the technology and we can say 'yes we have scoured the market and we are reasonably sure'. It is a matter of how efficient you can be with the resources that you are given."

Still early days

Artificial IntelligenceBut it's all still early days yet. David explains there are a variety of products that are either add-ins or very basic AI and we’re still a way from maturing.

"I think there are two challenges - one is the technology itself being matured and embedded into current technology. I think that will take a few more years worth of work - it's going to be an evolving process. The greater challenge is really about acceptance and behavioural changes inside of HR departments and it’s becoming clear in large corporates that technology changes have outstripped their ability to implement it inside the organisations," David says.

In the old days, he continues, the best technology was inside the office. Today, people's phones are generally more powerful than most PCs. "That's the world we live in. You've also got a bunch of senior executives who are now in their fifties - the baby boomers - who grew up in a world where the PC didn't exist when they started work. Now the pace of technology has outstripped their ability to keep up with some of the changes.”

David says a big phrase currently is many HR departments are 'data rich' but 'insight poor'. The theory behind the technologies has been around for years in terms of the mathematics, but it's only now that computing powers have caught up with how that can be used. "An example is that you can predict attrition in a business unit and then you can start modelling; if you give a five to ten per cent pay rise or bonus, will it actually keep those people? Next, it will start to predict that...or whether additional training will give that additional that's where it's going."

But in the end, even if 'computer says no', as the saying goes, according to David, the human factor cannot be eliminated from the process. "I think while machines can make the links in the data, it will still be human beings that will make the massive intuitive leaps between sets of circumstances," he concludes.