top of page

Humans invented AI - the future of recruitment is ours but augmented with technology

Recruitment is not dead. Humans created the robots. Technology is under our control and is there to help.

people meeting a new recruit

So much has been written and debated about the future of work,

robots displacing humans, Artificial Intelligence and then the possibility of technological singularity (when artificial intelligence is so advanced that it renders humans obsolete) that I think that we at times forget that we humans programmed the robots in the first place and that we can also stop them, slow them down or direct them to have a useful purpose.

It seems that recruiter bashing is the flavour of the day much like estate agent bashing was the trend before technology changed their profession as well. So as a former lawyer and headhunter it seems I represent two industries which are set for some big changes. I thought maybe we should look at actually what elements of recruitment we can automate, what elements create efficiency and where humans are still more useful than a bot.

Recruiters are not just trying to look out for themselves - I think everyone sees the writing on the wall to some extent. If you had asked me or any group of headhunters whether linkedin would have been a success 15 years ago we probably would have all laughed out loud. Why? Because simply putting your name, work details and availability in an open forum was the antithesis of what candidates wanted back then. How wrong I have been.

Just after the recession hit in the mid noughties I saw that the world changed and with a few others - Brave New Talent being the most hyped of all - we started to build peer to peer communities. Mine was twosteps, a Saas recruitment platform, with behavioural profiling. According to our clients, mainly large multi national corporates and professional services firms, they had enough of paying recruiter fees and wanted something else. I started that process nearly ten years ago and as much fun as it was it did not work. The reason being we could not get enough traction with high qualified and highly skilled professionals in permanent roles and our client's innovation budgets were then tiny. As soon as the market rebounded they defaulted to type and starting paying recruiters again.

The seed had been sown though and we may have just been a little too early. So what has changed now.

Well certainly technology is far more advanced and AI is proving to work better and smarter. I am looking forward to what Laszlo Bock and Humu are going to come up with. French AI recruitment tech firm Clustree just raised $7.9m on top of the $2.9m they raised in 2015 and at first glance they have some interesting products being used by the likes of Orange, Carrefour and L'Oreal.

So let's look at the core functions of recruitment, who does them and whether the human element can be replaced.

1. Research & mapping. Predictive analytics.

In order to find talent we need to know who they are and what they are doing. Tom Bowles a computer scientist was reported in 2016 to have been commissioned by The Founders Factory in London - think Brent Hoberman of lastminute and - to develop AI that works as a recruiter and he spent a year writing the algorithms. The aim he says was to remove human bias and the software has a significant impact on diversity because it does not care about background just the ability to do a job. The software can manage all of the research that a human does, enters it into a database and produces reports. It looks at how long people have stayed in role, promotions, work at fast growing companies, overseas stints etc. This is similar to what Clustree does.

What a headhunter does though is something a little beyond this - having conversations with people to reference and research who their peers rate as exceptional talent and how they perform. Humans can look at nuances in character.

Can AI detect the high performing egomaniac or the leader who sucks everything out of his team and loses them pretty quickly? The answer is yes if you are looking at internal data but possibly not if you look externally. The problem being that AI is data driven and you need to find good data to start defining patterns. Data on linkedin and other public systems is only as good as the human who entered it but you may see patterns in employer history.

Primary research, cross referencing and data input can be conducted by technology which is much faster, more accurate and cheaper than a human. That research is then passed onto a human to quality check and to engage that person in a conversation. If that conversation is with a employer will it be as frank as with a headhunter? That remains to be seen. The place of a headhunter or external recruiter has always been to create a layer of confidentiality and neutrality between the candidate and the employer.

In fact Clustree's tag line is 'Focus on Human, we take care of data'. So where we get to with research is that technology can read data far better than a human can and see patterns which can form hiring trends. AI is also used to predict patterns and trends using cold hard data whereas humans require longer to analyse trends but also rely on intuition.

AI removes bias at all levels. It can even identify the best candidates from a pool of cv's and search externally for relevant candidates from existing databases. What it cannot do is find those candidates who want to remain passive and reasonably secret. Humans build relationships with candidates. AI sees patterns and acts on them.

A significant area of human fallibility is in database management. Recruiters are humans and like many salespeople they don't like having to input data as they see this as eating into their sales time. However, having an incorrectly coded candidate, client or poor record keeping can create more risk financially or reputationally than pretty much anything else. So if the data is entered at the outset by a bot then we should be able to rely on the fact that it has performed all of its duties. Bots cannot be lazy. Humans can then add to or modify the data.

2. Screening

Having identified candidates the next job is for a recruiter to talk to them. AI may be able to define the perfect interview questions based again on data driven analysis and even predict the perfect set of questions for each role and then analyse the responses.

What AI will not be able to do, at least for now, is to read energy and body language. To decide that an interviewer needs to take a different path in order to get the best possible result is a human trait and we still need humans to execute the interview. Humans like talking to humans.

3. The interview stages.

Much has been written about the ability of technology to ensure that the interview process is short and efficient by scheduling and issuing prompts to follow up. This is one key area where candidates can lose interest especially if there is more than one employer chasing a candidate and the interviewer lets the process slide.

This is an important stage which requires finesse and that still means human interaction. The format for the interview may of course change with video interviewing, video cv's etc but I cannot imagine a situation where a human will not be on the other side of the table. What may happen is that we find that we have one less human interviewing and a bot feeding questions and analysing the responses far quicker than we can but that is perhaps a while away.

4. Negotiation of the offer and acceptance.

While technology may well be able to determine the exact remuneration package for a candidate, the skill of a recruiter or external headhunter is in getting the candidate over the line. Ensuring that offer is competitive and one that the candidate will want to accept.

This is not about making a placement fee or keeping to a KPI set by your employer, this is where the human element really comes into play. A good recruiter will not influence the candidate beyond providing market intelligence and information and having a thorough understanding of their needs and weighing this up with the offer. By this I mean being persuasive but not manipulative. There is little point in convincing a candidate to join a business which they then leave quickly as the fit was wrong.

Perhaps it is simpler to think of it like this, the human is the mentor, counsellor and adviser whereas AI provides insight, trends and information far more efficiently than any human can and without any bias at all. The latter part is of course provided that bias has not been built into the algorithm either overtly or covertly.

My last point is that as with everything to do with technology, it is only as good as the person who has programmed it. You may have the best systems in the world but if they are coded improperly they are worse than useless - again that comes down to humans.

The recruitment industry and the way in which we find and retain talent will of course change and for the better, we humans still have a significant part to play in it. Though beware the mavericks as if you have not fully grasped what it is to be human then your days are numbered at least career wise.

bottom of page