Follow zdnet: Please add it as a priority source On Google.
Important takeouts from Zdnet:
- Only 5% of enterprise customers benefit from generated AI.
- A bottom-up and top-down approach can improve the success of implementation.
- AI companies have made big promises in the bubble, but most of them are not fulfilled.
Investing in Generating AI may be booming, but most individual businesses using it have yet to see the payoff. in fact, New MIT research 95% of companies looking to leverage technology found that revenue and growth were not measurable.
Also, according to Gartner's 2025 Hype Cycle Report, AI Disillusionment General is coming
Research conducted by MIT Network Agents and Distributed AI (NANDA) The project was based on interviews with over 150 business leaders and analysis of 300 business developments in Generated AI.
“Only 5% of integrated AI pilots are drawing millions of value, while the majority are stuck without the impact of measurable P&L,” the author wrote in the report.
It depicts the total contrast between promise and reality. While technology developers sell AI tools like agents as productivity boosters, Nanda's new report shows that for all but most minorities, the technology has little or no impact on company revenue. How do you explain the huge gap?
What's not working – and what can you do
It is summarised primarily in the issue of bureaucratic inefficiency. Generating AI tools can provide increased efficiency in the hands of competent individuals, but as business leaders try to integrate them into existing company-wide operations and workflows, they tend to throw wrenches at organizational machines.
Also: 71% of Americans fear that AI will “too many people do their jobs forever.”
The main reason for this is that generative AI systems that most companies are trying to deploy internally and at large scale do not have the ability to seamlessly adapt to existing organizational workflows, ultimately causing more obstacles than acceleration.
“The central barrier to scaling is not infrastructure, regulation, or talent. We are learning,” the author writes. “Most Genai systems do not retain feedback, adapt to context, or improve over time.” The ability to remember past interactions, customize output to different contexts, and learn over time, are all important traits of AI, but the author specifically refers to the context of technology use within enterprise-scale operations.
Therefore, one of the implications of new research is that, in contrast to a top-down approach, it is bottom-up (which allows employees to experiment and discover the optimal mode of human and collaboration) in order to make the most of generative AI (forces all employees to use certain tools in a way that is closely controlled by executives and supervisors).
Also: Stop using AI for these 9 working tasks – here's the reason
Another trend that emerged from this study was flawed in prioritization in the application of generative AI. Many companies that were unable to profit from technology used it for marketing and sales, but the 5% who used it tended to do so through the automation of more subtle, mediocre “back-office” tasks.
Based on their research, the authors predict that future success will belong to businesses that deploy adaptive models with agents in the right place, and those who choose a general top-down approach will remain frustrating.
“The next wave of adoption is not the flashiest model, but “wins by systems that you learn and remember, and/or systems that are custom built for a particular process.”
AI hype and cultural pressure
On its surface, Nanda's work appears to support the belief that generative AI is nothing more than a massive hype bubble that quickly pops, unlike the short-lived corporate rush to the metaverse that precedes it. If most companies like this aren't seeing the results, that certainly means that technology is pedaling with an empty promise, right?
I know the time. For now, all-out companies are doubling their investment in AI, and they promise that the rise of more agent systems will mark a golden age of prosperity, creativity and leisure. At the same time, and soon after the mixed reviewed GPT-5 launch – Openai CEO Sam Altman himself He said he saw the AI bubble in shape..
Also: 5 ways automation can speed up your daily workflow – easy to implement
On the other hand, the broad cultural embrace of AI means that businesses are facing great pressure to quickly integrate technology, or at risk of looking dinosaurs. As Nanda's research shows, this rush appears to be done in many cases at the expense of all sorts of well-calculated plans, and as a result, investment in generated AI leads many companies anywhere.
Even at individual levels, generative AI can be counterproductive in the long run, even if it increases current productivity. For example, a recent study conducted by Workday found that there is a correlation between heavy AI use in the workplace and employee burnout.
