Consulting Is Ending Because Its Forecasts Carried No Information. So Do The AI Ones
The same certainty attached to the metaverse and to cloud computing, and it is attached to artificial intelligence now.
Sign up for my Substack daily AI newsletter here.
See my AI Training course portfolio for corporate Business Leaders here.
Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/
I run an AI advisory practice, so an argument that a rival form of advice is finished should be read with that in mind. What follows is not a case against management consulting. It is a case against the forecasts that industry produced, and against the identical forecasts now being used to justify AI spending in South African boardrooms.
CONTEXT AND BACKGROUND
Writing in TechCentral in August, the technology veteran Jannie van Zyl argued that management consulting as practised for half a century is finished, undone by the technology its own practitioners have been overhyping. His central charge is not that the firms were wrong about the metaverse. It is that they were exactly as confident about the failures as about the genuine shifts, and that their certainty about the metaverse was indistinguishable from their certainty about cloud computing.
The record supports him. In October 2022, McKinsey put the metaverse at up to five trillion dollars by 2030 and reported that one hundred and seventy-seven billion had already been invested since 2021. Citi projected between eight and thirteen trillion. Gartner forecast that by 2026 a quarter of the world would spend at least an hour a day in the metaverse, and that thirty per cent of organisations would have products ready for it.
INSIGHT AND ANALYSIS
It is 2026. None of that happened.
What makes this instructive rather than merely embarrassing is that someone did get it right, in public, at the same conference where those numbers were being repeated. The analyst firm Canalys called the metaverse a solution looking for a business problem and predicted most corporate projects would be closed by 2025. That analysis was available to any executive who wanted it. It simply came without a trillion-dollar number attached, which made it commercially useless to anyone selling readiness work.
That is the mechanism worth understanding. A forecast in the trillions, dated far enough ahead that nobody will check, creates the anxiety that the follow-on engagement resolves. The number is not an analytical output. It is the opening move in a sales process, and the firms that produce it face no consequence when it fails, because by then they have moved to the next wave.
The evidence is still on the internet. McKinsey’s report remains published, describing the metaverse as too big for companies to ignore, with the potential to generate up to five trillion dollars by 2030. No correction has been issued. No revision has been appended. The page simply sits there, four years on, and the same institution now publishes numbers about artificial intelligence.
Van Zyl’s phrase for this is that a signal firing identically for the true and the false is not a signal. He is right, and the implication is uncomfortable for anyone currently building a business case. AI is plainly more consequential than the metaverse. That does not make the forecast any more informative, because the forecast was not informative when the underlying technology was real either.
IMPLICATIONS
For boards, the practical response is to stop treating market sizing as evidence. A projection of what an industry might be worth in five years tells you nothing about whether a specific investment in your organisation will return. Those are different questions, and only one of them is answerable.
Van Zyl offers a test worth adopting. Ask the person pitching to define the central idea without the buzzword or the trillion-dollar number. Ask what they were certain of three years ago and how it aged. Then propose tying the fee to the outcome.
I should be candid that the third question is uncomfortable for my own trade as well. Outcome-linked advisory fees are resisted across the industry, including by people selling AI advice, and the reasons are not entirely unreasonable, since advisers rarely control implementation. The question remains worth asking precisely because of how a room responds to it.
For executives specifically, the useful discipline is to notice who was right last time. The sceptical view was on the record, from a named firm, at the same events. It attracted no follow-on work, because being correct and being sellable are not the same thing.
CLOSING TAKEAWAY
The strongest argument against management consulting was never that its forecasts were wrong. Forecasting is difficult and being wrong is forgivable. The argument is that the confidence carried no information, and that being wrong carried no cost.
Neither condition has changed. The institutions producing the AI numbers are the same, the method is the same, and no one has yet been held to a metaverse projection. Boards approving AI capital on the strength of market sizing from these sources have not learned the lesson they believe they have learned. They have simply changed the noun.
Author Bio: Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net
Comments