Can A.I. and the web predict your soft-skills?

People claim all kinds of soft-skills on their resumes. They find themselves to be very creative, with excellent English skills, killer presentation capabilities, while being a top problem-solver and of course a self-motivating team-worker...

How much of that is true? Recent studies with recruiting firms show that 53% of resumes contain lies and errors about their skills.

Which is why it is not very surprising that 91% of recruiters manually check candidates' online footprint to verify their claims from their CV.


But is the assessment of Soft-Skills really important?

Yes say 4,000 recruiters and HR professionals interviewed for the Linkedin Recruiting Trends Survey 2017, where Soft-Skills Assessment was placed as 2nd most important trend for 2017.

Also our own informal survey in 2016 with executives from Stepstone, Freelancer, Hays, Adecco, Randstad and Xing revealed the same: assessing soft-skills is a key need for the recruiting industry.

Why is that? To quote a CEO of a German Mittelstand firm:

"We only know how to assess hard-skills but we mis-hired 100 people last year who weren't able to communicate well, present to clients or work in a team."



Recruiters and Hiring Managers use an array of tools to measure soft-skills:

1) Resume

The good old resume is on its dying leg, with over half of them containing lies and errors. Even if correct, they only give a mini glance at someone and will never reveal enough information to assess hard- and soft-skills of a candidate.

2) Tests

There are many online tests that predict personality and soft-skills. The main (and huge) drawback is that candidates have to take and re-take them again and again. In a recruiting world that calls itself as being in a 'war for talent', especially in the IT-skills, this is not a tool that a talented engineer wants to go through 4 times a week.

3) Interview

The 'Gold-Standard' of skills assessment is now often replaced with video interviews, sometimes already with pre-recorded video interviews. While such pre-recorded videos are great in terms of time-savers for candidates, they really show a well-rehearsed side of them.

The in-person or live interview, is of course still how people like to make decisions with their 'gut'. Having hired over 50 people already in my career though, I learned the hard way that relying on my gut from interviewing people isn't a good indicator of how they will behave in the job. I am sure there are great recruiters and HR people out there who can make great decisions about soft-skills from interview, but I think the vast majority of hiring managers (in SMEs), get influenced by people's showmen skills and interviewing skills.

4) Assessment Center

Using Assessment Centers to role-play roles in teams and observing candidates over a day isn't a bad idea, but they are very costly and time intensive for candidate and hiring firm.

From my own experience, I took part in 2 assessment centers in the summer of 2000 for the same firm: Accenture. The first was for an internship and I failed the Assessment Center miserably because I behaved like myself - which was apparently too much of a bully for Accenture (ooops). In the second one just 2 weeks later for a full-time job, I just sat back and contributed meekly. And not being met got me the job offer...

5) Reference Calls

Checking people's skills in calls with former employers (should) be standard practice. But even if companies actually do them, such calls often just reveal glowing, generalized statements as most employers are just happy that person is off their payroll.

5) Online Footprint

With every day and week there is more and more online content about people on the web. Our presentations, tweets, likes, answers, patents, jobs, they all become more and more transparent and a real true measure of our skills.

Especially the analysis of our online footprint for indicating our soft-kills. We can analyse your English skills from your public content. We can determine if you are really a problem solver by your activity on Q&A sites, competitions or patent sites.

In a famous study from Cambridge University on using Facebook 'likes' to predict personality, their A.I. was able to more "accurately predict the person's personality than a work colleague by analysing just ten Likes; more than a friend or a cohabitant (roommate) with 70, a family member (parent, sibling) with 150, and a spouse with 300 Likes."

That is the power of AI combined with the web footprint of candidates. Its not scary, its actually the fairest way to assess skills.

* * *

By Simon Schneider