BCG's new North America chair counted his own technical staff and found twice as many

Mel Wolfgang became chair of Boston Consulting Group's North America business in May and asked a question most executives assume is already settled: how many of the firm's traditional consulting staff had skill sets resembling those of its technical employees ^[1][2]. His answer, as he told Business Insider, was roughly double the workforce he thought he had ^[3], which makes this less a story about AI adoption than about how poorly capability is tracked inside large firms.

According to Wolfgang, nearly as many people with traditional consulting backgrounds, including what he called "the English lit major with an MBA," had developed skills approximating those of BCG's technical hires ^[4]. The firm arrived at that through tests and employees' demonstrated technical work on the job ^[5]. Two caveats sit inside the finding. It is a self-reported internal audit disclosed in an interview, and the operative word is "approximated" ^[4]; the source does not state what threshold separates a consultant who codes from a software engineer. Both caveats matter, because the number is only useful if the categories are defensible.

The commercial context makes the audit awkward rather than reassuring. Wolfgang said BCG is accelerating hiring of both technical talent and business generalists ^[6], asks candidates in interviews whether they have used AI tools such as Claude and Codex ^[7], and is looking for the "Venn diagram overlap" of technical proficiency and critical thinking ^[8]. A firm that has just discovered it owns twice the technical bench it assumed is, at the same time, buying more of it. That is the pattern worth naming for any operator approving a hiring plan this quarter: the requisition is often a proxy for an inventory nobody has taken.

The training data shows what happens after the inventory. BCG's AI fluency curriculum is structured as a four-step ladder that includes programming a tool using multiple agents, with assessments that adapt when a staffer is struggling ^[9][10]. About half of the North America workforce has completed level one, about a third level two, and about 20% level three ^[11]. Completion at level three is therefore roughly 40% of completion at level one ^[12]. The stated goal is all staff through level one by year-end and all four levels within a couple of years ^[13], which means roughly doubling level-one completion in the time remaining ^[14]. Wolfgang, who said he recently finished tier two ^[15], wants people to pause between steps and apply the material, comparing the written driving test to the road test that happens on casework ^[16].

He is explicit that the tooling does not carry the work: someone still has to frame the problem, work the analytics, and judge whether the answer makes sense ^[17], and initial AI output may need several passes and substantial refinement ^[18]. Wolfgang called judgment "an apprenticed skill" transmitted by officers and project managers sitting with junior staff ^[19][20], and said critical thinking remains "in the bull's-eye" of what the firm trains ^[21]. He put 30% to 40% of BCG's work as AI-focused, with the rest traditional strategy, mergers and post-merger integration ^[22][23] - implying 60% to 70% is still conventional consulting ^[24] - while adding that close to 100% of engagements touch AI in some way ^[25].

Watch whether level-one completion actually reaches 100% by December, whether the skills audit changes the hiring mix or merely sits alongside it, and whether BCG ever publishes the definition of "technical" that produced the doubling.