BCG Is Turning Its Generalists Into AI-Fluent Consultants

BCG's Push to Blend Technical and Judgment Skills

At Boston Consulting Group, the consultant of the future increasingly looks like a new kind of generalist. Mel Wolfgang, who became North America chair in May, says traditional consulting staff are now developing technical capabilities that were once concentrated among specialists in areas such as IT architecture and software development. After surveying the firm's conventional consultants, Wolfgang said he found that nearly as many had skills approximating those of BCG's technical hires.

The discovery is now shaping hiring. BCG is accelerating recruitment of both technical talent and business generalists, but Wolfgang says the firm is searching for the 'Venn diagram overlap' between strong technical proficiency and the critical thinking consultants have traditionally needed. Job interviews now ask about experience using AI tools such as Claude and Codex. Even so, Wolfgang argues that technical ability is not enough: consultants still have to frame clients' problems, work the analytics and decide whether an answer makes sense.

AI is not a narrow practice area at the firm. Wolfgang said roughly 30% to 40% of BCG's work is focused on AI, while 'pretty close to 100%' of engagements now touch on AI in some way. In response, BCG has expanded training beyond new hires and people entering new client industries. Consultants and support staff learn the basics of building AI agents and choosing where to use them, and employees who want deeper skills can study data engineering.

The firm's four-step AI fluency program includes technical tasks, such as programming an AI tool that uses multiple agents, and forces participants to work through judgment questions. Assessments adapt if someone is struggling. In North America, about half of workers have finished the first level, about one-third the second and about 20% the third. BCG aims to have all staff complete the first level by the end of the year and all four levels over the next couple of years.

What BCG's AI Training Reveals About the New Consulting Model

Wolfgang Is Redefining the Generalist Without Abandoning It

Wolfgang's survey challenges the assumption that technical capability must be bought through specialist hires. His claim that traditional consulting staff — including, as he put it, 'the English lit major with an MBA' — often match the skills of technical employees suggests BCG already has a larger AI-capable workforce than its org chart shows. The strategic value is clear: if the firm can validate and mobilise those hidden skills, it can serve clients' AI needs without expanding its specialist hiring at the same pace. However, this is still an internal assessment by a senior executive, and the source does not show how those skills translate into billable AI delivery.

AI Is Now a Layer Across All of BCG's Work

The 30% to 40% share of AI-focused work is only part of the story. Wolfgang says nearly every engagement now touches AI in some form, meaning even traditional strategy, merger and post-merger work now includes an AI dimension. That changes the minimum technical baseline for almost everyone on a case team. It also means the firm cannot solve the problem by isolating AI expertise in a separate practice; delivery teams across the portfolio need enough fluency to assess AI outputs and ask the right questions.

A Four-Step Training Ladder Becomes the Delivery Engine

BCG's response is an assessed, adaptive training program that is explicitly meant to be applied on casework. Wolfgang compares passing a training tier to passing a written driving test; the real test, he says, happens on client work. That design is significant because it treats judgment as an apprenticed skill rather than a credential. The completion figures — half at level one, one-third at level two, 20% at level three in North America — show substantial but uneven progress. The main execution risk is whether completing training tiers translates into consistently better client outcomes, especially since Wolfgang says early AI output often requires multiple rounds of refinement.

BCG's Competitive Logic Is Talent Leverage, Not Just Tool Adoption

BCG's move signals a shift in how professional-services firms may compete for AI work. By training business generalists to use AI tools and encouraging hires who already know Claude or Codex, BCG is trying to combine low-cost scaling of AI capability with the premium it attaches to judgment. If that model works, the firm can offer AI services without being wholly dependent on a separate technical bench. If it does not, clients may see inconsistent depth across teams. The source does not provide data on client results, so the outcome remains an open question.

What BCG's Shift Means for Consulting Leaders, Hires and Clients

The story carries concrete signals for consulting leaders, people seeking consulting roles and clients buying AI-related advice.

• For consulting firm leaders: Before hiring externally for AI capacity, test the technical skills of existing business generalists. BCG's review found nearly as many traditional consultants with technical-hire-level skills as it had expected, suggesting hidden internal capacity may already exist.
• For consultants and job candidates: Expect interviews at firms following BCG's lead to ask about hands-on AI tool experience, including Claude and Codex. A credible answer should pair tool use with problem framing, analytics and the ability to judge whether an AI-generated answer is actually sound.
• For professional-services employees: BCG's four-step ladder — starting with AI-agent basics and building toward data engineering and multi-agent programming — is a useful structure for applied AI training. Its key design is a pause between levels to apply the skill on real work, which BCG compares to passing the road test after the written exam.
• For clients buying consulting work: Ask whether the team serving you has demonstrated AI fluency or only completed a course. Nearly all of BCG's engagements now touch AI, but delivery depth varies: about half of North American staff have finished only the first training level so far.