84% of AI projects failing entirely or failing to scale, and 64% of CEO's still throwing money at them.

Pretty obvious what part of the company needs to be replaced.

https://fortune.com/2025/05/09/klarna-ai-humans-return-on-investment/

As Klarna flips from AI-first to hiring people again, a new landmark survey reveals most AI projects fail to deliver

Just 1 in 4 AI investments bring in the ROI they promise—but CEOs just can’t resist the technology.

Fortune

@johnefrancis the one that keeps opening up the door to every tech company that shows up and says "bright shiny objects? I've got a whole bag of them right here!

"You want a bright shiny object? That's a good boy, now fetch!"

@johnefrancis Someone was pitching us to have AI generate quotes.

My first question was, so they pay if their tool misquotes?

@johnefrancis "Sorry folks, but we're going to have to get you all to work longer hours because the stock price took a big hit last quarter. There isn't enough money for any raises, either - it all went to the CEO's golden parachute. If only there were some way we could have seen this coming!!"

🤦‍♂️

@johnefrancis They need to get Evan Solomon on the job.
@mpjgregoire this is a serious edit from Avi Lewis to Evan Solomon!
@johnefrancis the tech industry is fuelled by peer pressure. Less leadership and more followship (sic).
@johnefrancis exactly. I work in research where its being used more yet we haven’t studied AI effectively, don’t know what it’s costing us and should reconsider https://open.substack.com/pub/alteragents/p/the-future-of-insights-isnt-ai?r=1osp1&utm_medium=ios
The Future of Insights isn't AI

by Devora Rogers, CSO, Alter Agents

"Agents of Change" Substack by Alter Agents
@johnefrancis I've said this before, but I wish I could work a job where I could so abjectly fail at my prime prerogatives and still get promoted, transferred, or retire wealthy early.

@johnefrancis We are already transitioning into the phase of disillusionment. Much money will have been put down the drains, but some of the tech will live on and make a marginal difference to some companies.

It will not revolutionize industries, products or services. Here and there new products and services will have been created, some improved, and some will have gotten worse from using this new shiny tech.

Only rarely do we encounter new technologies that fundamentally change how we do things.

Steam Engine, Atomic Energy, Internet.

I personally have not seen LLMs to exhibit even remotely similar effects on society as these other inventions have had.

It would be fun to be wrong though.

@nopatience @johnefrancis But... but... but...

AI was never meant to be the revolution... and Agentic AI was *actually* going to bring about the real change.

And then when Agentic AI is found to provide much less than the hype promised, there'll be another flavour that was always going to be the promised child.

@nopatience @johnefrancis I always considered LLMs and NFTs to be in the same category of uselessness honestly. LLMs are just taking just a bit longer to die off

@USBTypeSteve I will say, for the record, that I find it to have it's uses... but my use-cases are pretty old and well-established (categeorization and summarization).

Pretty well-understood problems.

But the whole idea of having companies run by "agent" employees... I would LOVE to see this happening. Begin with replacing the CEO, because when it comes to bullshit and hallucinations... AI-agents are pretty solid.

@johnefrancis

CEOs are herd animals. They follow the leader without knowing anything about the direction they're headed.
#AI

@johnefrancis Hopefully this is a good sign in general, hopefully the people that make so much money they get detached from reality will finally learn what we always knew.

@johnefrancis

ANYONE who has used LLM's (it's not AI, it's fancy autocomplete) for anything knows it's not all people cracked it up to be.

The code it writes is crappy. The information it gets from stealing information from other sources isn't accurate and that's just the start.

Look at OpenAI, the "premiere" LLM company --- they've been on the edge of bankruptcy for years.

@johnefrancis CEOs and managers are using AI for their own work and getting away with it.

Blowing up a single-sentence email to 5 paragraphs to look professional. Generating summaries of emails because nobody wants to read more than one paragraph.

@johnefrancis The money boys. The boards. The lawyers and veeps who decided this was the future, and not a black box that'll decimate their staff and alienate their customers.
@johnefrancis but all of their friends are jumping out of the proverbial bridge. /s

@johnefrancis @Gargron

We all love to hate AI, but that stat? I'm surprised Fortune doesn't know that 2/3rds of ALL tech projects fail? And for largely the same reason.

Take a stroll through one of those "tech innovation hubs" where VCs troll for profit. So many are just such an obvious 'Nope' but they won't stop until the money runs out.

@johnefrancis Perhaps AI business churn is one of those human activities that doesn't need to produce anything to be profitable. Like professional sports or something.

@johnefrancis

Some of the most impactful decisions are made by people who don’t understand what they are dealing with

@johnefrancis "After months of boasting that AI has let it drop its employee count by over a thousand.."

And there it is...the core problem. Every product and service is made for and ABOUT people. When you replace PEOPLE (with anything, llm's, automation, etc...), you've already gone against what your product/service is all about.

It's almost like companies are thinking "we're all about people...but not our employees, they're different...they're just an expense".

@johnefrancis

It'll be interesting (and sad) to see how much of the 500 billion dollars or whatever the US is going to throw at the AI industry is just going to go down the drain as spectacular fiascos, environmental catastrophe, and also human catastrophe, as people are mislead and forced by circumstance into industry careers that lead them to an unfulfilling life and future

@johnefrancis This statistic isn't useful without comparing to the percentage of non-AI tech projects that fail similarly.
@johnefrancis
I've been making this point for a while now.
The most obvious part of most companies that could easily be replaced with LLMs is senior executives.
But somehow this argument isn't popular with the people making the decisions where money is spent. They prefer "developers" or "marketing people" or "creative artists" as targets
@johnefrancis expecting AI to take over from people is like expecting a parrot to take phone calls. They can politely say good morning and a few other phrases but lack the comprehension to do the job.
@johnefrancis clearly it’s all the employees failing to embrace the plagiarism machines that are making more work and problems for them that’s why they have to lay so many off

> McDonald’s tried an AI-driven system for its drive-thrus for three years before ending the effort. During those years, the system made mistakes like trying to add bacon to an ice cream order and giving one customer an order of 260 Chicken McNuggets.

lol

@johnefrancis No. Those are the numbers the CEOs self-reported. The results of AI projects are actually far worse, these are only the ones they couldn‘t ignore
@johnefrancis
Fuckin truu, i clap to that🙌🙌
@Gargron