Hot Tip: That ChatBot on Your Apply Process Isn't AI

By Linda Brenner | January 26, 2024

For all the TA and HR leaders who have invested in these technologies, I implore you to just try it. You might be surprised to find that the Bot on your application page is actually repelling the talent you most want to attract.

It's easy and quick to test: Try typing in any of the following and play the role of a highly qualified candidate for one of the jobs you'd most like to fill:

"I'm a software engineer and interested in finding out more about your open roles."

"Can I confirm my interview time?"

"Do you hire interns?"

"I'm an ICU RN moving to (XX city) and would like to talk to someone about your open positions."

"I'm having a hard time completing the application. Can someone help me?"

"What data analyst jobs do you currently have open?"

What's wrong with these Bots? A lot:

They operate within a very narrow scope. In most cases, they can only loop back incessently to information in the "open jobs" list. This is not information that candidates could possibly know, so it almost always sets them up for a frustrating experience.
They have limited comprehension. For example, if one uses abbreviations (CPA, RN, Ops, etc.), the systems often cannot comprehend the request.
They don't "learn" from user interactions. As we all know from using ChatGPT, true AI constantly learns, adapts and improves on its responses. But these hiring ChatBots are static and insufferably un-smart. It's almost impossible not to fall into a small, circular loop of information that does not address a basic question.
In the worst example of "not learning", if you click away from the chat (say, to look at the job ads it suggests), your previous chat isn't saved; the user is forced to and to start all over again. In the worst of these cases, they even require the user to confirm their acceptance of their privacy terms over and over again.
Many Bot chats never even request the user's contact information which in theory could be used by recruiters to follow up with the candidates that appear to be the most valuable. This is the worst of all misses with these systems because, bruh come on - that doesn't even require AI.
Most of these systems don't have the "intelligence" you'd truly want to identify or alter the experience for a software engineer or maintenance technician or seasoned RN candidate from roles for which there is ample talent available.

Just imagine you led TA for a tech company and a candidate asked your Bot about a critical opening - say, a software engineer role - and the Bot couldn't quite respond in the appropriate way. Wouldn't it be great if it would respond with: "You've maxed out my abilities. Could I get your contact info and connect you to a recruiter right away?" Then, it would notify a recruiter who would *immediately* contact this potential hot candidate. Because - let's face it - you're not going to overwhelmed with software engineering candidates looking to talk to a recruiter. Just that basic work stream could be a TA game changer for in-bound talent.

What's a company to do?

If you have one already, test it. Demand better. If you don't have one but are considering it, try before you buy by testing them on multiple customers' platforms (it may take 10 minutes). You might be shocked to find how unhelpful (repellant?) they actually are. Certainly, this technology will improve in the future but until then, our advice: proceed with great caution.

Examples:

These includes examples from the two largest Chat companies and across several clients . . . see for yourself (note that often, the system only allows one to choose from selections such as "corporate" vs "field" - which often don't even make sense to people outside the organization, but I digress. In these cases, the user cannot type anything - only choose from the selections offered):