Posted on November 27, 2018

AI in Recruitment – Now your entire team knows more than your most experienced recruiter

Part of an Ideas in Action showcase presented by Tijs van Tilburg at the SIA Executive Forum 2018.

Tijs van Tilburg, our Business Development Director for Staffing & Recruitment Solutions, recently presented an Ideas in Action showcase at the Staffing Industry Analysts Executive Forum in London. During the presentation, Tijs showed how your entire team can leverage the knowledge and understanding of your most experienced recruiter through the power of AI-powered solutions. The below is an excerpt of Tijs’ presentation, you can read on or watch the full video at the end of the article.

Understanding, analyzing and connecting people and jobs

What we have previously been teaching people in recruitment is to understand how to talk to the machine. How to talk to the machine in the way that the machine will actually execute what you want.And, I’m not sure if you have applied that method in your Staffing or Recruitment business, but if you have you may agree that it is labor-intensive to get people trained to that level. And, there are a lot of people that excel at it, actually so much that that’s what they’ll do for the rest of their lives. They’ll call themselves Boolean black belts and may become a freelancer that teachers other recruiters Boolean. But not everyone does that. And a lot of staffing and recruitment consultants hate it. So, what they will do is copy something someone else made and reuse it. It’s very hard to maintain. It’s very hard to get that skill at a certain level, and to keep it at that level throughout your business.

In recent years we’ve see the rise of what we call semantic search, which turns that topic around. Instead of teaching you how to talk to the machine, we want the machine to be able to understand you.So how do we do that? One way is to teach it to understand what you type. If you type in something, you want to use your natural language, right? You don’t want to use brackets, OR statements or other command line coding. You just want to type what you’re looking for. And you may make mistakes. And you may actually have a different meaning to what you’re typing. So, it makes a difference if you’re looking for someone that has an accountant certification, or somebody that is currently a certified accountant. That context makes a difference in what you’re trying to achieve.The second way is you want the machine to understand what you mean, not what you type. If you type “CPA”, you don’t want everything that includes the letters CPA, you want “Certified Public Accountants”. You want people that are what you mean when you type CPA. And specifically, in multi-lingual environments, you want to express that question in your language, be it English or Dutch or German, but you may have people in your database that have a profile in a different language. So, you might actually want to find people that have a German profile but are exactly what you mean when you type CPA – even if they don’t use the same language, words or even letters.

Another thing people in recruitment know, is that what you’re looking for is far more nuanced than the combination of all the words you type in to your system. And you’ll see that because the reason we actually articulate our questions is often in very long job descriptions, so if we need that much text to actually tell people what we’re looking for, what we expect of them, there is some very important nuance there.One of the nuances is: “How important is this?” Do you need to speak Spanish, or is it preferred? Would you be preferred over other candidates if you speak Spanish? And what level of Spanish do you need? Is intermediate enough? Or do you need to be advanced or native? The same thing goes for experience. If I’m looking for a Java developer that knows Scrum, I may actually prefer people that actually have more than 4 years of Scrum experience. So, understanding the granularity between level, importance and proficiency is something very important to be able to express.

One of the reasons many people, even if they have a lot of tools, will still express their questions in title, location, and maybe education or skill is because often it’s very hard to actually articulate this question. Many people in our business, when we ask them “Hey, can you articulate the question of what you’re looking for exactly?” will have a hard time actually giving you an answer” So understand and recognizing relevance makes it a lot easier to come up with an answer.

What if I have 80 people that have a CPA in London with a Bachelor’s or a Master’s degree, who speak Spanish? How do I recognize that maybe people that who have experience with Sage or Quickbooks or similar are probably more relevant for me or for my customer than other people? Having the data insight/visualization show “Hey, of these candidates that you have now, this many people live closer than the max radius” makes it a lot easier to dig deeper.

The last part is you want to curate knowledge. It’s very helpful if you get all of these suggestions, but again you want to create your own concepts; what you call ‘top prospects’ or what you call ‘high potential’ is very unique to you. You may actually be able to find information or be able to add information where you think “I want to narrow down or expand this definition of something”. And using that knowledge to actually improve the performance of everyone in your team is something very powerful. If your top recruiter understands that certain concepts are very important to find more relevant candidates in your area of focus, you want other people in your company to benefit from that. You want to curate that knowledge and retain it in the “machine”, even if that person leaves your organization.

If you’d like to know more about the concepts discussed here or more detail from Tijs’ presentation, please fill in the form below:

Contact us

  • You can find more information about how Textkernel collects, uses and processes your data in our Privacy Statement