This weekly update shows that artificial intelligence agents are changing how businesses work, and it is affecting millions of workers around the world. An AI agent is a computer program that can think, make choices, and do tasks all by itself without a person telling it what to do at every step. Companies are moving fast to use this technology because it can work 24 hours a day and make decisions very quickly.

The biggest news this week is that Amazon announced it is cutting 16,000 corporate jobs. Amazon says the company is doing this because AI can now do work that people have done for years. The company also explained that AI will do to office jobs what machines did to factory jobs a hundred years ago. Amazon made a lot of money recently, so the company is not cutting jobs because it needs to—it is cutting jobs because it believes AI can do the work cheaper and faster. Business expert Parag Amin says that many companies are excited about AI, but they are learning that AI still makes mistakes and needs people to check its work and fix problems.

Job losses from AI are happening in many countries, but the United Kingdom is being hit the hardest. A study found that British companies cut 8% of their workers because of AI in the last 12 months. This is twice the average for other countries like the United States, Germany, Japan, and Australia. Britain is having trouble because workers are expensive to hire right now, and government rules have raised the cost of paying workers. So companies are using AI instead of hiring people. In the U.S., companies are also cutting some jobs because of AI, but they are also creating new jobs in other areas, so the total number of jobs is staying more steady.

The most interesting trend is that agentic AI is moving from experiments to real work. Right now, about 40% of company software programs are expected to have task-specific AI agents built into them by the end of 2026, and these AI agents will do things like make sure cloud computer costs stay low, stop security problems before they happen, and check that money reports are correct. These AI agents will run automatically without waiting for people to approve every choice. Companies like McKinsey say that 40% of their total workforce is now made up of AI agents—about 20,000 AI workers out of 60,000 total.

However, there is a big problem nobody expected: much of the work that AI creates is broken or useless. This problem is called "workslop"—it means AI-created work that looks good on the surface but does not actually work. Companies are losing about $9 million dollars every year for every 10,000 employees because of this problem. The reason this happens is that 41% of employees say their bosses told them to use AI but did not explain how to use it correctly or why they should use it. When people use AI just to show they are following orders instead of to actually get better work done, it does not help the company.

Another problem is that employees are not really seeing benefits from AI, even though the technology is saving them time. A study of 3,200 workers found that people save up to seven hours every week using AI, but almost 40% of that time is spent fixing, rewriting, or double-checking the AI's mistakes. This means that only 14% of employees actually saved time in the real world after using AI. Workers also report that saving time with AI just means they get more work dumped on them—32% of employees said this. Companies are more likely to spend their AI savings on buying more technology (39%) than on teaching their workers (30%).

Business leaders who are succeeding with AI are doing something different: they are treating their workers like partners in change, not as people to replace. Research from KPMG found that the best-performing companies are the ones that teach workers new skills, build trust, and make workers feel like part of the AI change. These winning companies plan to hire more workers in 57% of cases, compared to only 35% for other companies. Only 6% of workers at these winning companies feel left behind by AI, compared to 39% at other companies. This teaches us that companies that invest in people win more than companies that just buy technology.

Interestingly, young workers have a different view of AI. Nearly 80% of people born after 1997 (called Gen Z) think AI is making their coworkers lazy and less able to think for themselves. Yet these same young workers are using AI even more than older workers. This shows that even people who worry about AI see it as something they need to use to keep up with their jobs.

Looking ahead, the big challenge for businesses is "machine identity security"—making sure companies know what all their AI agents are doing and keeping them under control. As companies use more AI agents, they need strong rules about what these agents can and cannot do, just like they have rules for human workers. The companies that will win in 2026 are the ones that treat AI agents like important workers that need careful management, not like simple tools that run by themselves.