By Ashkan Rajaee, Founder, TDZ Pro (tdzpro.com)
Can you imagine being interviewed by a bot?
Not a chatbot that asks for your email. A simulated client, talking back to you in real time, while you try to sound like the future of a firm you've dreamed about joining.
That's what McKinsey is piloting this fall. According to reporting from The Wall Street Journal on Oct 14, The Verge and the New York Post, some entry-level business-analyst candidates in North America will take a roughly 10-minute interview with an AI playing the client. One example: an AI mayor who wants recommendations on a stadium. Hundreds of candidates may use it. The AI doesn't score anyone. Human assessors in the room make the call. And McKinsey hasn't published any success metrics.
Those are the facts as reported. Everything below is my opinion.
Why this caught my eye
I've spent my career in software services, growing a consulting company from about $6 million to $60 million as a partner, then building a fully remote consultancy. I've hired a lot of people. And the first thing I notice here is not the technology. It's the direction of travel.
My read: this looks like another sign that larger companies are trying to reduce the cost of hiring by automating more of it. To be clear, McKinsey didn't say that. That's my inference, not their claim. Their stated aim is to see how candidates think and respond. But when you're evaluating hundreds of people, a simulated client that never gets tired starts to look a lot like a scalable process. You can see why that's attractive.
The candidate experience
Put yourself in that chair. You're 22, you've prepped for months, and your "client" is software.
Some people will love it. No nerves about impressing a partner, no small talk, a clean shot at showing how you think. Others will freeze, because talking to a bot feels strange and nobody's taught them how.
Here's what I wonder most. Real client conversations have texture. A client sighs, changes the subject, pushes back because their boss is in the room. Can a simulation capture that? Maybe. I don't know, and McKinsey hasn't shared data that would tell us.
Bot first, human second
Here's the part I keep coming back to. A candidate gets interviewed by a bot, then, if they get the job, takes direction from a human manager.
How receptive will people be to that sequence? Some will shrug. Some will wonder whether the firm values them as people or as data points to process. First impressions of an employer stick. If your first real "conversation" with a company is with a machine, what does that say about the relationship you're signing up for?
Who's behind the script?
This is the one that really gets me.
When a person who has worked at a company talks to you, you can weigh what they say. You know they've lived it. You can ask follow-ups and read their face.
When the AI plays the client, what's it been programmed to say? Candidates can't see the prompt. They don't know what the system rewards, what it's been told to push back on, or how its personality was set. It isn't the same as a human who's been in the room with real clients.
To be fair, the humans watching are the ones deciding. That's a real safeguard. But the candidate's performance is shaped by a script they never see. That deserves a lot more transparency than it's gotten so far.
What I'd want to see
If companies are going to use tools like this, I'd want three things:
- Tell candidates up front exactly how the simulation works and who decides.
- Publish results. Does it predict who succeeds on the job? Right now there are no published metrics.
- Give candidates a human alternative.
None of this is anti-technology. I just think hiring is a place where the burden of proof should be higher, because the people on the other side are betting their careers.
Your turn
I'm genuinely curious where people land on this.
What are your thoughts on being interviewed by a bot?
Disclaimer: This article reflects the author's personal opinion only. It does not represent the views of any company named, and it is not a statement of fact or legal advice.
Top comments (0)