This as-told-to essay is based on a conversation with Jason Richard Desentz, Toshiba’s chief human resources officer for the Americas. Desentz has been in the HR industry for nearly 30 years and joined Toshiba more than two years ago. The essay has been edited for length and clarity.
When I first got into HR, not everyone had a computer. People had notepads and paper in front of their desks, and they still worked.
Now people are asking: How can I become a recruiter if an AI automation goes through the résumés? How am I going to learn to screen a candidate and move up the career ladder?
I don’t approach this as an AI story. It’s an adaptability story, and AI is another chapter.
Having AI doesn’t mean I won’t still have junior-level employees. The junior job will just be different, and I will be hiring differently, too.
Junior recruiters can become consultants sooner
For a lower-level job at a company like Toshiba, a junior-level HR could face a thousand résumés. Am I actually going to go through the full thousand? No. I would be here all day, seven days a week, reading résumés.
Typically, you’ll go through the first 50, and if you get 10 good résumés, you’ll stop. You could be missing a gem in the rest.
We’ve built AI bots, including one for recruiting. The aim is to surface stronger candidates and free recruiters to enhance other parts of the process.
Instead of reading résumés one by one, they can focus more on the intake: How can I make my questions for managers better?
I also want to put more effort into onboarding and that first-day magical experience: Employees walk in, have everything they need, and are trained.
We use AI to create interview guides from a résumé, job description, and manager’s intake form. These can be personalized to the candidate, identifying areas where we should ask more questions without eliminating someone who may be able to learn.
Even in a junior role, your job becomes more of a consultant rather than a transactional, “Here’s a résumé.”
I ask how you think more than what you can do
Rather than saying, “Can you build a pivot table in Excel?” I’m looking for people who are really good at problem-solving.
You may not have work experience yet. Tell me about a time in college when you were presented with a problem. How did you solve it? Walk me through your mental process.
That gives me insight into your critical thinking. Do you pull the problem apart? Do you try to understand what’s causing it?
I’m looking for data literacy, too. Have you used information in other areas of your life to defend a position?
Then I’ll ask about a group project. I’ll follow up with, “I bet you had someone on your team that didn’t pull their weight.” How did you get that person motivated? How did you influence them to contribute? They’re going to get the same grade you get, whether they put in the work or not.
I also look for comfort with ambiguity. You come in with a presentation ready, and someone gives you information that knocks it all down. How do you handle that?
One scenario I might give is designing a recruitment process with a budget. Halfway through, the CEO says we have to cut 20%. How are you going to reconfigure your process?
Or I’ll ask: You had time to finish a project, but you got sick. Now you only have one day. How did you get it done? What did you do differently?
I shift the interview toward capabilities and away from specific skills. Nothing is static in life. Everything is dynamic.
Take the mystery out of AI, and give people time to play
The AI fear is real: the fear of missing out and the fear of being obsolete.
As a leader, don’t dismiss it. Instead of saying, “AI won’t take your job,” I would say, “The work is changing. We’re going to navigate this together.”
When I met with my team to build AI bots, I asked: “What’s the No. 1 thing you don’t like about your job?” People were bogged down with simple questions. Let’s get to that low-hanging fruit first.
At Toshiba, we have about six courses covering AI basics. We’ve had about 500 course enrollments and 300 completions.
You’ve got to give people time to play. We experimented with about 12 use cases. Some bots were still initial experiments.
We also did a little “Shark Tank,” where employees presented their bots and how they built them. The winner got a small prize.
Have fun with it. It’s OK to make mistakes. You don’t learn how to ride a bicycle the first time.
