Is Tier-1 Consulting Hiring the Wrong Talent, Again?

Tier-1 consulting firms like McKinsey, Bain, and BCG have a very specific operating model.

They send teams into a client site to solve hard problems fast.
The consultants don’t know the company “from head to toe.”
They’re often in their 20s or early 30s.
And yet they’re expected to walk in, absorb chaos, and produce clarity.

So for the past 50 years, the hiring logic has been consistent:
If you want people who can survive ambiguity at speed, you hire the smartest, fastest learners you can find.
Elite undergrads. Top MBAs. People who have proven they can win in competitive environments.

And consulting firms built screening machines to find them.

One example was McKinsey’s old written test: the Problem Solving Test (PST).

If you never experienced it, imagine this:

A room full of GPA 4.0 students.
They already made it through resume screening.
Everyone is “top of the class” somewhere.

Now you hand them a one-hour multiple-choice test.
No domain knowledge required.
Just reading, logic, math, and speed.

And then the brutal part:
Only 20% of that room moves on.
Most people walk out, even though on paper they look like future partners.

The test was consulting in miniature:

• Read a long prompt

• Decide what matters

• Ignore what doesn’t

• Pull insights from scattered paragraphs

• Do fast calculations

• React quickly

Then firms started to question whether “book-smart under multiple-choice pressure” predicts real-world performance.
Real problems are dynamic. The process matters. There isn’t always a single right answer.

So, they evolved.

PST turned into what people call the problem-solving game: more like a simulation.

You’re dropped into scenarios.
You must understand the goal, use what you have, and adapt as conditions change.
The way you adapt becomes the evaluation.

Now here’s what’s been bothering me:

When you look at the PST-style test today, top AI models would destroy it.
They can read 100 pages in minutes, extract the core arguments, and structure the answer cleanly.
What takes a human 60 minutes can take AI five.

And when you look at the “game” version, AI agents are moving in that direction too.
They can operate through a GUI, observe what’s on screen, reason about next steps, adjust, and iterate.

So here’s the question:
If Tier-1 consulting firms built recruiting around skills that can increasingly be replicated by AI tools that cost $20/month… what’s left?

And if what’s left is the real value…
Are they hiring for the wrong things again?

At NitroLens AI, we believe what’s left for strategy consultants is human.
The most important information isn’t in Snowflake or Salesforce.
It lives in people’s heads, scattered across teams.
AI can analyze what’s written. It can’t extract what isn’t.

Follow us — we’ll share more POV on what truly creates value in strategic projects in an AI-native world.