If you’re applying for a job these days, there’s a growing chance that your first interview won’t be with a human. The initial weeding out of job candidates is yet another area into which artificial intelligence is beginning to venture.

The commercial real estate giant JLL hasn’t gone that far, but it’s nevertheless using AI as a tool for improving workforce hiring, recruiting and efficiency, especially for the procurement function.

AI is already proving its mettle at sorting through stacks of resumes to find the candidates who best match the company’s hiring requirements, says chief procurement officer Michael Raphael.

The success of the tool, of course, depends entirely on the quality of information that’s fed into it. Raphael says JLL’s hiring managers first must build the specific criteria on which the AI model will rely, in order for it to spot the most promising candidates.

What AI isn’t doing — at least for now — is stepping in and conducting JLL’s initial job interviews. The latest iteration of generative AI is perfectly capable of doing so, although most job candidates would likely argue that the results fall short of fully replicating human speech and interaction.

The question to consider, Raphael says: “Are we taking something that should be a personal experience and making it impersonal?” And, given that JLL isn’t filling thousands of positions each week, AI’s ability to cull in that way isn’t required.

What the technology can do is help equip JLL’s client account teams with suitable individuals, as part of its outsourcing service, as well as select the best candidates for its extensive intern program. With AI, Raphael says, “we’re making sure that the CV matches with what we want.”

JLL is very much invested in developing AI internally, having created its own in-house version of ChatGPT. And the system’s value is extending well beyond the recruiting and hiring function. “It’s a tool that we can put our own proprietary information into, and feel that confidential information isn’t going outside the firm,” Raphael says.

A possible tipping point is when the tool becomes so effective in responding to internal queries that it eliminates the need for some human hires altogether. One of JLL’s clients, for example, has an AI-created receptionist. The benefit of a virtual agent, Raphael says, is that “in all situations, it’s going to respond the way you want it to.”

For now, though, the chief role of AI at JLL remains that of an internal support mechanism for humans. Raphael says it serves as a mentoring tool, giving recent hires access to extensive information on procurement policy and procedures that would otherwise have to be conveyed by company veterans. “I want Michael as CPO to be able to get a similar experience to Suzy who joined two days ago, and doesn’t know anybody,” he says.

Think of it as a modern-day incarnation of the “expert systems” envisioned by scientists in the earliest days of AI development. But Raphael cautions that JLL GPT still has a way to go before matching the knowledge base of people, and conveying that information in the most effective manner.

One of the system’s shortcomings, he says, is that many of its responses can be frustratingly verbose. “All of the content is real,” Raphael says, “but is it relevant to me as an individual? Does repeating things three different ways add value?”

So, can an individual applying for work at JLL forever be assured that its initial interaction with the company will be with a human, not an AI bot? Never say never, Raphael suggests.

“Perhaps as things progress, we may start using AI for the interview process,” he says. In any case, he adds, the experience of being grilled by AI might not seem so creepy to a younger generation that’s accustomed to existing in a virtual world — with “characters” that previously have been confined to video games.