Accenture, the consulting firm, built a reputation as too big to fail. In June 2026 it had its worst day in years on the stock market, even though revenue and earnings per share grew that quarter. Investors sold anyway, pricing in what AI might do to future demand for consulting, and analysts pointed to nothing specific to the company. Close to a fifth of its value disappeared in hours, part of a slide of more than 50 percent in its share price that year.
The sell-off traces back to what consulting firms sell. Accenture alone has more than 700,000 employees billing hours across Fortune 500 boardrooms. Firms like it built revenue on sending teams to help companies roll out new software, then billing for months of setup and training. That software has come to run well with fewer people managing it, so clients need fewer outside teams. A firm built on advising companies about new technology finds itself advising them on how to run with fewer advisors.
AI companies have also started sending their own engineers into client companies. These small teams, called forward deployed engineers, build the software themselves instead of handing over recommendations, and they stay until it runs on its own. Hundreds of consultants once spent a year on work that a handful of these engineers can finish in weeks. Two AI companies, OpenAI and Anthropic, have set up separate units that work this way inside banks, insurers and other large companies.
Clients are spending less too. Saudi Arabia’s Public Investment Fund, the country’s sovereign wealth fund, sharply cut its use of outside advisors in 2026. Boston Consulting Group, which built one of its biggest markets around the fund, has taken the deepest hit. It stands to lose hundreds of millions of dollars in revenue from that one client. McKinsey and PwC’s strategy arm have lost work too, though less.
Both pressures landed within the same few months, yet bookings at large consulting firms have not collapsed. Executives at firms including Accenture argue that AI creates new advisory work as it removes old tasks. For the largest firms, the test is whether they can build new services around smaller teams before the shift outpaces them.
