WHO IS consulting good for? Consultants, obviously. Chief executives, who can blame failure on bad outside advice and take credit for successful counsel. Also, for the industry’s one listed behemoth, its shareholders. Between the start of 2015 and the end of 2024 Accenture, which split off from its accounting sibling in 2000 and went public a year later, generated a total return (including dividends) of around 370%, handily outdoing not just the S&P 500 index but also Goldman Sachs and Morgan Stanley, rival redoubts of advisory smugness. As America’s stockmarket climbed to an all-time high in February, the firm was worth $250bn, more than either investment bank.

Since then, however, investors have wiped some $60bn from its market value—and a self-satisfied smile off its face. On June 20th its share price tumbled by 7% following a disappointing quarterly earnings report. Revenue and operating profit both rose a touch faster than expected year on year, to $17.7bn and $3bn, respectively. American-government contracts took less of a hit than feared from the DOGE efficiency drive. But new bookings declined for the second quarter on the trot. Both one-off consulting projects and “managed services”, where Accenture runs certain corporate functions day to day on clients’ behalf, were down. The number of individual customers inking over $100m-worth of business with the firm in the previous three months dipped from 32 to 30.

Some of this is a temporary setback. Amid the fog of a trade world war, and of geopolitical ructions in the Middle East, many global companies are currently preoccupied with survival rather than “reinvention”, which is Accenture’s stock-in-trade. Still, the firm’s problems run deeper. Having made a fortune telling others how to adapt to newfangled tech, from the internet to cloud computing, it now faces the selfsame predicament in the age of generative artificial intelligence. As semi-autonomous gen-AI “agents” sweep the world, who needs consultants?

This is the uncomfortable question before Julie Sweet, Accenture’s boss since 2019. Her two-part answer has been to insist that clients will require as much help with gen AI as they did with earlier tech innovations, or more, and that Accenture is perfectly placed to provide it. Neither claim sounds persuasive.

It is true that plenty of multinationals can make neither head nor tail of gen AI. Ask most managers about the relative virtues of Claude Sonnet 4 and ChatGPT o3 and you get a blank stare. A recent survey by S&P Global, a data compiler, found that 42% of companies abandoned most of their AI initiatives. A year ago the figure was just 17%. Clearly, then, some hand-holding is in order.

But for how long? Accenture’s success was built on partnerships with a plethora of technology providers, whose often finicky products it has long helped clients select, put in place and maintain. All sides stress the strength of their enduring relationships. In November, for instance, Accenture and Microsoft added a “Copilot business transformation practice” to Avanade, their 25-year-old joint venture. In May the consultancy and SAP, a giant of enterprise software and another longtime collaborator, unveiled a new programme to help small but fast-growing companies “reinvent, thrive and grow” (in Ms Sweet’s words) and “move faster, operate more efficiently and scale with confidence” (in those of Christian Klein, her opposite number at SAP).

For all such public bonhomie, though, some of Accenture’s partners cannot wait to cut out the middle man. AI is being integrated into their offerings so that it works straight out of the box—and keeps working as AI agents automatically update and upgrade IT systems in accordance with users’ commands. Newcomers like Palantir are embedding their own engineers with customers. All this lets clients save money on Accenture consultants, in the blunt words of one supposedly friendly tech boss.

Already the pace of Accenture’s new gen-AI contracts is slowing, from $200m a quarter last year to $100m in the past three months—not exactly reassuring for what are the early days of a ballyhooed technological revolution. It implies that, for Accenture, “AI is not digital 2.0,” sums up Tom Rodenhauser of Kennedy Intelligence, which tracks the consulting industry.

Despite Ms Sweet’s insistence to the contrary, the AI age looks likely to belong not to enablers of technology like Accenture but to its originators. Consider the past decade. In the seven and a half years before ChatGPT introduced gen AI to the masses in November 2022, Accenture’s shareholder returns, of 200% over the period, and its future prospects, as measured by the ratio of its share price to forecast earnings, dwarfed those of companies such as SAP and IBM. In the two and a half years since, the tables have turned. Palantir, for its part, is worth $338bn, six times what it was just a year ago.

Accenture could have used its access to capital markets to invest in deep tech (which IBM, for example, has continued to do despite a pivot to consulting in the 1990s). Instead it opted to splurge on innumerable “tuck-in” takeovers of small consultancies. That includes maybe 50 ad and marketing agencies that, if Meta and Google have their way, gen AI is about to make obsolete.

In an attempt to calm investors’ nerves, Ms Sweet has reorganised her firm around “reinvention services”. The new unit combines all of Accenture’s businesses into a one-stop shop to meet clients’ needs. It will be run by Manish Sharma, the well-regarded boss of Accenture’s American operations. This sounds an awful lot like, well, Accenture—and Mr Sharma’s new role like Ms Sweet’s old one overseeing the whole business. If the firm really wants to avoid being disrupted out of existence by AI, it may need some better advice. ■

Subscribers to The Economist can sign up to our Opinion newsletter, which brings together the best of our leaders, columns, guest essays and reader correspondence.