How to interview at McKinsey, BCG and Bain in the age of AI

Frame the problem, and aim high

The first minutes of a case test whether you can pick the right question. Models are good at answering the question they’re given. Picking it, and deciding how ambitious the answer should be, is still a person’s job. Asked at CES what the models can’t do, McKinsey’s global managing partner answered:

In an interview, that means restating the client’s goal, asking the one question that could change your plan, and saying what a great outcome would look like before you reach for a framework.

Make the call, and stand behind it

Analysis is getting cheap. Deciding what it means for a client, and saying so to their face, isn’t. Ron Kermisch, who leads talent acquisition at Bain, puts it plainly:

Faster analysis raises the bar rather than lowering it. Anna Mattsson, a senior partner at McKinsey, says “what that actually means is that we require judgment, much better judgment” (Anna Mattsson, Business Insider (YouTube), opens the source). In a case, judgment is a recommendation made before you have every number: the answer first, the reasons, the main risk, and what would change your mind.

Check the numbers yourself

Recruiters now expect candidates to be fluent with AI tools. BCG’s CEO, Christoph Schweizer, told the Wall Street Journal: “We are blown away about the junior talent we can hire. They are so AI native, they use these tools as if the world had never existed without them.” (Christoph Schweizer, The Wall Street Journal (licensed reprint on bcg.com), opens the source) The basics haven’t moved, though. Alicia Pittman, BCG’s global people chair:

A study of 758 BCG consultants shows what can happen when people stop checking. On a business problem the AI handled badly, the researchers found that “for a task outside the frontier, subjects using AI were 19 percentage points less likely to produce correct solutions” (Dell’Acqua et al., Organization Science, opens the source). Reading an exhibit is the interview version of that check: the units, the base and what’s missing, before you trust the headline.

Learn fast, and learn from people

Leaders at the firms say they hire for how fast you learn, not for what you already know about AI. Christin Owings, who leads BCG’s London office, told the FT: “We also continue to hire people who might have almost zero knowledge of AI when they join, because we hire people who have deep curiosity.” (Christin Owings, Financial Times, opens the source) Tamar Dor-Ner, Bain’s Americas regional managing partner, says “it's probably even more important that you're built to keep up” (Tamar Dor-Ner, Bain’s Beyond the Bio podcast, opens the source).

Much of that learning still happens person to person. Bain’s CEO, Christophe De Vusser:

A market you’ve never sized is a small test of this. There’s no pattern to match, only reasoning you can explain out loud and correct as you go.

Recover, and be yourself

McKinsey’s leader has said the firm used to screen differently. Bob Sternfels told HBR’s IdeaCast that it had been “too focused on, “Did you have perfect marks?” versus, “Did you have a setback and recover?”” (Bob Sternfels, HBR IdeaCast, opens the source) The firm also wants the person, not the script. Marie Christine Padberg, who co-leads McKinsey’s global talent attraction:

The experience interview tests both. Pick a real setback, say what you changed, and tell it the way you would to a colleague rather than reciting it.