On 13 October, The Guardian reported that companies that rushed into AI are now pulling back . Meta scaled back plans to cut staff and use more AI agents after employees spent more and more time cleaning up technical and security errors, citing Reuters . Walmart store and warehouse workers described errors and false alerts from AI tools that assign tasks and schedules, citing Business Insider .
At Microsoft, executive vice president Ryan Roslansky wrote that he received an AI-generated document, had his own AI summarise it and ended up with nothing new: “If we’re not careful, we’ll end up with a very expensive way to avoid writing and reading in an ocean of sameness, genericness”
Questions this article answers
SAQ: the questions you should be asking (and nobody is)
Most companies got no faster with AI.
The video for this article is in production.
What the numbers say
Here is the data that exists. The Guardian, citing an Atlanta Fed working paper, reported that 90% of executives say AI efforts have yet to translate into productivity gains at their companies . The Atlanta Fed paper “Firm Data on AI”, a survey of almost 6,000 executives in the US, UK, Germany and Australia, reports “more than 80 percent of firms reporting no impact on either employment or productivity” over the last three years . One counts executives, the other counts firms. Both point the same way.
The same paper found that firms expect gains over the next three years, forecasting a 1.4 percent productivity boost . So the honest reading is not “AI does nothing”. It is “most firms have not seen it yet”.
Why rushed rollouts stall
Johnny C. Taylor Jr., CEO of the Society for Human Resource Management, told The Guardian that leaders feel “we’re driving the truck and building it at the same time, and that’s a problem” . Microsoft’s chief strategy and transformation officer wrote:
We learned that access and usage do not equal transformation
Kathleen Hogan, Microsoft
Wharton’s Jessica Reif told The Guardian that AI tends to produce “plausible-looking output” whose mistakes are not immediately apparent, so errors can spread through a workflow . Meta also told employees that AI usage would no longer be part of performance reviews, The Information reported, as cited by The Guardian .
What the firms that gain do differently
Microsoft says its sales team stopped pushing adoption and started from business goals. Within that group, according to Microsoft’s internal data, adoption of priority use cases tripled, revenue per account manager rose 9.4% and close rates were 20% higher . The company also wrote that people “set the requirements, correct what falls short, determine where human judgement is essential and remain accountable for how AI is applied” .
The business lesson
Sources
Sources are listed in their original language.
- 1The Guardian, Taylor Telford, “Companies rushed in on AI. Now some are pulling back,” 13 October 2026. https://www.theguardian.com/technology/2026/oct/13/ai-companies-rushed-workers-fallout
- 2Ryan Roslansky, LinkedIn post, “AI slop isn’t happening just on social networks…”, 2026. https://www.linkedin.com/posts/ryanroslansky_ai-slop-isnt-happening-just-on-social-networks-activity-7496622259214659585-Kxjf
- 3Federal Reserve Bank of Atlanta, Yotzov, Barrero, Bloom et al., “Firm Data on AI,” Working Paper 2026-3, 24 March 2026. https://www.atlantafed.org/research-and-data/publications/working-papers/2026/03/24/03-firm-data-on-ai
- 4Microsoft, Kathleen Hogan, “What we’ve learned from Microsoft’s own AI transformation,” The Official Microsoft Blog, 17 September 2026. https://blogs.microsoft.com/blog/2026/09/17/what-weve-learned-from-microsofts-own-ai-transformation/