BCG earns US$ 3.6 billion from AI while McKinsey cuts 10%: consultancy against itself
BCG generated US$ 3.6 billion from AI work by 2025; McKinsey plans to cut 10% of its staff. The firms that sell automation to the world have begun applying the tool against their own profit pyramid.
The Boston Consulting Group generated US$ 3.6 billion from work related to artificial intelligence in 2025, which is a quarter of its US$ 14.4 billion revenue, according to figures revealed by Bloomberg on April 23. Just days earlier, a plan reached McKinsey's partners to trim approximately 10% of its workforce in roles that do not serve clients, affecting a few thousand people over the next 18 to 24 months. These two announcements describe the same industry. Within it lies the contradiction that will characterise consultancy in 2026: the firms that profit most from selling automation are the first to turn the tool against their own payroll.
The money is in sales, cuts are backstage
Neither of these firms is shrinking due to a lack of demand for AI. Accenture accumulated US$ 5.1 billion in generative AI bookings in the twelve months leading to the end of 2025 and increased its team of data and AI specialists to around 77,000 people, up from 40,000 two years prior. Accenture itself has decided to stop reporting AI figures separately from the second fiscal quarter onward; CEO Julie Sweet argues that the technology is now embedded in almost everything the firm delivers, making the isolated metric meaningless. When the indicator becomes noise because AI is ubiquitous, the boundary between "AI project" and "consultancy" ceases to exist.
The cuts, for now, are targeting elsewhere. The approximately 200 job terminations that McKinsey made at the end of 2025 impacted internal technology and support, not the workforce that charges client hours. Nearly half of the firm's staff are in back-office roles, precisely where generative AI today performs in minutes the research, scheduling, and formatting tasks that junior analysts used to take weeks. The message from Bob Sternfels, the global managing partner, was explicit: further reductions in non-client-facing roles in the next two years, while the firm promises to hire more frontline consultants.
The pyramid sustaining profit is cracking
The classic consultancy model is a pyramid: many junior analysts at the base produce the material that a few partners at the top sell at a high price. AI precisely attacks the base. If an analysis that previously occupied a team of recent graduates for days can now be produced in hours, the firm charges for fewer hours, and the mechanism that transformed cheap labour into high margin loses strength. McKinsey's revenue has remained practically stagnant between US$ 15 billion and US$ 16 billion over the last five years, following a phase of aggressive hiring. Generating the same revenue with fewer people is the polite definition of a cut.
There is a counterpoint that tempers the apocalyptic script. The job cuts have not yet affected the personnel serving clients, and McKinsey states it will hire more in this area. Gartner and Forrester have modelled that 20% to 30% of revenue from large strategy firms comes from AI, and BCG's 25% falls within this range, not above it. The money from AI is real, but it still competes for the budget with traditional restructuring work. The pyramid has not collapsed: it is being redesigned with fewer steps at the base.
OpenAI and Google push the chips to the centre of the table
The bet has become non-optional. On February 23, OpenAI partnered with BCG, McKinsey, Accenture, and Capgemini to sell and implement its Frontier agent platform, with engineers from OpenAI allocated within the projects and teams of the certified consultancies. BCG and McKinsey come in as strategy partners, while Accenture and Capgemini act as end-to-end integrators. Intuit, State Farm, Thermo Fisher, and Uber are already appearing as initial clients. In April, Google Cloud announced a US$ 750 million fund to support consultancies in adopting agent-based AI, according to Bloomberg.
For those leading an IT consultancy in Brazil, this movement serves as a dual-warning. The same AI that compresses billing for the giants is also compressing yours, and the alliance between OpenAI and the Big Four diminishes the advantage of those who merely resold technical capacity. Value remains for those who provide what the agent cannot deliver alone: business context, accountability for results, and the decision of where automation should and should not be applied. The firms that wrote the manual for others’ efficiency now need to prove that they know how to apply it without draining their own profit engine.