AI Hot Air

Posted February 26th, 2019

By Sergei Makhmodov, founding director, DaXtra Technologies

The current noise revolving around AI hype is deafening. Every field and industry is touting artificial intelligence, from finance and banking to retail and education. The recruitment industry has not been left out of the brouhaha.

The terms "AI", "machine learning" (ML) and "deep learning" (DL) are all being increasingly overused and overhyped by marketing and sales professionals of numerous recruitment technology vendors and often, by ill-informed end-customers themselves. 

I am sorry to disappoint you, but most of it is hot air.

Why? Well, because any "intelligent" computer system requires vast amounts of data to be trained on. Also required is significant computational power to process that data. So, how can a company with few referenceable clients and whose product has only seen a few hundred thousand documents at most, deliver anything more than a wasted investment?

Why is there so much clatter about AI, you may ask? The answer is simple: It sells! Unaware end-customers buy into overhyped and overpromised sales pitches and savvy marketing spiels, often without any proper thought, evaluation or objective comparison. 

Do not fall into this trap. Protect your investment by spending a little extra time on testing across large data sets and always research on the vendor and their solution before you buy.

Now, there are obviously many technological marvels out there, and RecTech is by no means an exception. But, end-customers need to educate themselves on what the current evolutionary stage of AI is, what can be expected from it and how to select the right solution to address specific business needs. Here are some important questions you should be asking when considering any new technology that is claimed to be "intelligent":

  • How many live, happy customers does this vendor have?
  • How long have those customers been using the software and what have the returns on their investments been?
  • What is the clients' attrition rate and what are their success stories?
  • Is the technology proprietary or from a third party?
  • How much real-life data has the technology seen and been trained on?
  • Who are the main people that developed the technology and do they have sufficiently advanced
    degrees in AI?

AI Daxtra Stats graphic

DaXtra currently processes upwards of 100 million free-form, unstructured resumes and job descriptions each month in more than 40 languages, through large distributed networks of its cloud-based servers. That is 1.2 billion documents a year that our software constantly learns from to literally understand complex unstructured texts, extracting statistical knowledge and inferences. With that in mind, is Daxtra’s technology AI-based? Very much so. Although we often downplay this fact, encouraging customers to test our products and judge by results instead. At the end of the day, what does it matter to the end-user, as long as expectations are met?

There's also a lot of talk and debate on whether AI will soon be able to automate the entire recruitment process. We certainly don't see it happening any time soon. Technology is just not 'human' enough to be able to intuitively assess, negotiate, seduce and close — all of which are qualities intrinsic to a good recruitment professional. The real question should be whether AI, in its current best form, is capable of helping recruiters do their jobs better, quicker and more efficiently. This goes without saying!

The key here is to be smart. You don’t have to be especially tech savvy to ask pertinent questions to get the answers you need in order to evaluate and make a wise decision when purchasing anything that is purportedly AI-based.

Here is a related article on this topic, written by DaXtra's CTO, Steve Finch - "Are robots really taking over? Untangling AI hype from reality".

Tags: AI, DaXtra Blog, Machine Learning, Deep Learning