The same firms selling Fortune 500s on AI transformation are now bleeding talent to the people trying to put guardrails on it.
The Summary
- Former consultants from BCG, McKinsey, and Accenture are leaving to work in AI safety, governance, and policy — staffing the UK AI Safety Institute, RAND, EU institutions, and policy teams at frontier labs
- AI safety roles now account for 40% of high-impact career transitions tracked by Consultants for Impact, double the share moving into global health
- The irony: consulting firms are remaking themselves around AI while their own people join the institutions trying to regulate what they're building
The Signal
Consulting has always been a talent farm for tech. The pattern is predictable: learn the frameworks, build the slides, then jump to Google or a Series B startup for equity and autonomy. But this new path is different. These aren't consultants chasing IPO windfalls. They're joining underfunded government institutes and think tanks to figure out how to keep AI from running off the rails.
Sarah Pomeranz's nonprofit, Consultants for Impact, tracks where ex-consultants go when they want their work to matter. The data tells the story: 4 in 10 high-impact moves now land in AI safety, governance, or policy. That's twice the share going into global health and development, which used to dominate the "I want to do good" exit route. Former BCG and McKinsey analysts are now writing regulatory frameworks at the EU, staffing the UK AI Safety Institute, and embedding in policy teams at OpenAI and Anthropic.
"A lot of these ex-BCG and McKinsey consultants are staffing the institutions that are, in many ways, governing."
This is happening while their old employers go all-in on AI. McKinsey has plastered generative AI across every practice area. BCG launched its own AI agent platform. Accenture is retraining armies of consultants to sell AI transformation. The firms smell revenue, and they're chasing it hard. But the consultants who actually built those decks are looking at what's being sold and deciding someone needs to pump the brakes.
The skill transfer makes sense. Consultants know how to synthesize complex systems, translate technical jargon for executives, and build frameworks that institutions can actually use. Those are exactly the skills needed to turn abstract AI risk into concrete policy. The UK AI Safety Institute doesn't need more ML researchers. It needs people who can explain why a banking AI that hallucinates credit scores is a systemic risk, then draft legislation that won't get gutted in committee.
Key factors driving the shift:
- AI safety organizations need operational muscle, not just technical expertise
- Consulting teaches systems thinking and stakeholder management — exactly what's missing in most AI policy work
- The "impact" career path now has institutional backing and clearer job pipelines
What's unclear is whether this is a trickle or the start of a flood. Namaan Mian from Management Consulted says the data hasn't caught up yet. These moves are still anecdotal, not statistically significant. But the fact that a dedicated nonprofit exists to facilitate this transition suggests there's enough momentum to matter.
The Implication
If you're in consulting and wondering where to point your career, this is a real option now. The institutions trying to govern AI are hungry for people who can build frameworks, not just write papers. The pay is lower than Google, but the leverage is higher than another deck about supply chain optimization.
For everyone else, pay attention to who's staffing these safety institutes. The people writing AI regulation in 2027 are the same people who were selling AI transformation in 2024. That's either reassuring or terrifying, depending on whether you think consulting teaches systems thinking or just PowerPoint.