Follow ZDNET: Add us as a favorite source On Google.
Key takeaways of zdnet:
- Only 5% of enterprise customers are making profits from generic AI.
- A bottom-up versus top-down approach can improve the success of implementation.
- AI companies are making big promises in a bubble, most of which are incomplete.
Investments in generic AI may be boom, but most individual businesses using it have not yet made payment. In fact, A new MIT study It was found that 95% of enterprises trying to exploit technology are not looking at the average result in revenue or increase.
Also: According to Gartner’s 2025 Hype Cycle Report, Gen AI disillusionment
Study, operated by MIT Network agent and decentralized AI (Nand) The project was based on an analysis of interviews with over 150 business leaders and 300 commercial deployment of generic AI.
The authors wrote in the report, “Only 5% of the integrated AI pilots are extracting millions, while the vast majority are stuck with no average P&L effect.”
This is a clear contrast between promises and reality: while tech developers are selling AI equipment like agents as productivity boosters, new reports by Nanda states that for all, but a missing small minority, technology has no effect on lower lines of businesses. What do you eat for heavy inequality?
What’s not working – what else can
It boils to a large extent in terms of bureaucracy disability. Generative AI equipment can provide efficiency benefits in the hands of competent persons, but when business leaders try to integrate them into existing, company-wide operations and workflows, they throw a wrench in the organizational machinery.
Also: 71% of Americans are afraid that AI ‘will make many people out of work permanently’
According to the report, the main reason for this is that the generative AI system that most businesses are trying to deploy internal and lack the ability to adapt basically with the existing organizational workflows on the scale, eventually make them more obstructed than an accelerator.
“The main obstacle for scaling is not infrastructure, regulation or talent. It is learning,” the author writes. “Most Genai system does not maintain reaction, adapt to references, or improve over time.” While the ability to remember previous interactions, customize the output in various contexts, and learn over time. All major symptoms of AI are, the author is specifically referring to the reference to the use of technology within enterprise-scal operations.
One of the implications of new studies appears to be the most generous AI for businesses, they do a bottom-up (they do well to use employees and to search their optimal mode of human-AI cooperation) contrary to a top-down approach (all employees tightly control us to use a special tool in a manner).
Also: Stop using AI for these 9 tasks – why is it here
Another tendency that emerged from study was a priority in the application of Generative AI. Many businesses that failed to profit from technology were using it for marketing and sales, while 5% who were using it were successfully prepared to do so through the automation of more nice and worldly “back-office” works.
Based on their study, authors estimate that future success will be related to businesses that deploy agent and adaptable models at the right locations, while those who choose a general, top-down approach will be disappointed.
“The next wave of adoption will not be won by the most attractive model,” they write, “but by those systems who learn and remember and/or by systems that are custom formed for a specific process.”
Ae promotion and cultural pressure
On its surface, Nanda study seems to support the belief that generic AI is nothing, but there is a large -scale promotional bubble that will soon pop, not the short -term corporate congestion, which was before it. If such a big ratio of businesses is not seeing the result, then it definitely means that technology is being pedaling on empty promises, right?
Only time will tell. For now, companies from all over the board are doubled on their investment in AI, promising customers and investors that the rise of more agent systems will enter the golden age of prosperity, creativity and holidays. At the same time-and a mixed review on a GPT-5 launch high heels-Openai CEO Sam Altman himself Said that he sees taking an AI bubble shape,
Also: 5 methods automation can speed up your daily workflow – and implementation is easy
Meanwhile, the extensive cultural embrace of AI means that companies are facing heavy pressure to integrate technology quickly – or risk like dinosaurs. As Nanda’s study indicates, this crowd, in many cases, is apparently being done at the cost of any well -calculated plan, and as a result, investments in generative AI are leading many companies.
Even at the individual level, the generative AI can be renovated long-term-even while promoting productivity at present. For example, in a recent study conducted by the workday, a relationship was found between the work and heavy use of AI in the employee burnout, while other studies found evidence that the use of AI reduces significant thinking skills.
