McKinsey deployed 12,000 AI agents. Here is what enterprise AI actually looks like at scale.

McKinsey has deployed 12,000 AI agents internally. One consulting team that previously required 14 consultants now runs with 2 to 3. The firm has publicly signalled plans to reduce headcount by 5,000 as AI takes on more analytical and delivery work.

The irony is precise. The firm charging $1,193 per hour for senior partners to advise enterprises on AI transformation is simultaneously replacing the people who deliver those engagements with AI agents.

There are two ways to read this. The cynical reading: McKinsey is reducing costs while maintaining fee structures, capturing the AI productivity gain for themselves rather than passing it to clients.

The honest reading: this is what enterprise AI at production scale actually looks like. Multi-agent orchestration replacing manual analytical processes. 14 people to 2 or 3. Not because the humans were doing bad work, but because AI agents can do the volume work faster, cheaper, and without the overhead.

The relevant question for enterprise CxOs is not "what is McKinsey doing internally?" It is "which of our own 14-person processes could run with 2 to 3 people, or fewer, if we built the right agent architecture?"

Weaver's 40 industry process maps are built around this question. O2C, P2P, RCM, claims lifecycle, content supply chain: processes that currently require 10 to 50 people, end-to-end, with multi-system orchestration. Every one of them is a candidate for the McKinsey reduction ratio.

The question is not whether your industry will be affected. The question is whether you are leading that reduction or reacting to it.