Leadership1 publisher2 min readPublished
Gartner's 30% and MIT's 95% count different populations of AI failure
Daniel Jabaraj's Forbes column pairs Gartner's abandonment forecast with MIT's zero-return finding to argue that AI programs fail on business framing. One study counts projects and the other counts companies.
The Board Room · Leadership desk
What happened
- Daniel Jabaraj, CEO of the developer tools company Syncfusion, argues in a Forbes Tech Council column that AI transitions fail on business framing, writing that technology is rarely the problem.
- He cites Gartner's prediction that at least 30% of generative AI projects will be abandoned after proof-of-concept because organizations cannot connect them to a clear business purpose.
- The column also cites MIT researchers who found 95% of organizations getting zero return from AI pilot programs, despite $30 billion to $40 billion in enterprise investment.
Compiled by The Board RoomSomething wrong?How this is made
Why it matters
- decision The only cause named in either citation can be tested at approval: a project that cannot state its business purpose fails Gartner's criterion before a dollar reaches a model provider.
- constraint A change-management line item cannot be defended with these two rates. They point at approach as the cause and put no figure on the remedy, so the cost of the fix has to be argued from a company's own numbers.
- exposure Both rates reach the reader secondhand, through a contributed column written by a software vendor's CEO, so anyone quoting them in a board paper is standing on methodology they have not read.
- precedent If a third of generative AI projects are expected to stop after proof-of-concept, an internal decision to kill one becomes an ordinary outcome that a sponsor can survive.
Gartner's 30% is a forecast about projects and MIT's 95% is a reported result about organizations, so the two figures are not two readings of one quantity [2][3]. Multiply the two headline numbers anyway and 95% of $30 billion to $40 billion gives $28.5 billion to $38 billion of investment sitting in organizations that report nothing back [10]. The step works only if every organization spent about the same amount, and the column does not say how that money is distributed [3].
The named cause is narrower than either percentage suggests. Gartner attributes abandonment to the organizations' inability to connect a project to a clear business purpose [2]. The MIT report, as Jabaraj describes it, says the delta between cost and success rates is not model quality or the technology itself but the approach [4]. Ownership, retrofitted process and training all sit inside that one word, and the citations here name only the missing business purpose.
The sequencing claim comes from a different source in the column. Jabaraj quotes a Harvard Business Review study of digital transformation failures which found that organizations "put the cart before the horse, focusing on a specific technology rather than doing the hard work of fitting the change into the overall business strategy first" [5].
Jabaraj is CEO of Syncfusion, which builds developer tools and enterprise software [1], and he wrote: "In my experience, technology is rarely the problem" [6]. A tools vendor has an interest in that sentence. The three citations came from elsewhere, though, and the argument he owns, that companies define themselves by method instead of mission and then have nowhere to go when the method is disrupted, rests on 25 years in the software industry [7][13].
For a leader deciding this quarter, the timing matters more than the percentage. The cause Gartner names is visible before a proof-of-concept begins, so the question of whether a project attaches to a stated business purpose can be asked at approval [2]. Jabaraj also writes that companies almost universally underestimate the work required for a real transition [8]. His first move for the exceptions is to set the mission at an altitude "broad enough to survive changes in method, yet specific enough to mean something" [11].
What to watch
- Whether the MIT report, read directly, supports the organization-level reading of the 95% figure that the column gives it.
- Any Gartner restatement of the 30% abandonment forecast as this year's proof-of-concept cohorts finish.
- A study that measures spending on ownership and process redesign against the return on AI pilots.