A new report shows that 95% of companies are getting no real value from their big investments in smart technology, creating a big gap between them and the few who succeed. The main problem is that most of these tools don't learn from feedback or remember past conversations, so they don't fit well into the day-to-day work people do. This is why most tests of these new tools fail. Interestingly, many employees are using their own personal tools like ChatGPT for work tasks and getting better results than with the expensive official company tools. The few companies that are successful are not building their own technology. Instead, they are buying special tools from partners that are made for very specific jobs. These tools learn and get better over time. These successful companies are finding that the biggest money savings come from back-office jobs, like paperwork and customer service, often by replacing outside agencies instead of firing their own staff. The future might be an "Agentic Web," where different smart tools can work together on their own to get complex jobs done.
MIT: Why Most Companies Fail With AI
Most companies get no value from new smart tools because the tools don't learn from their work.