Weekly Update: How AI Agents Are Changing Work Around the World

Big Companies Cut Jobs While Buying More AI

Companies across the United States are making a difficult choice. They are cutting jobs while at the same time spending huge amounts of money on AI agents. In 2025, companies in the US said they cut about 55,000 jobs because of AI. This is much more than before—it was only about 4,250 jobs two years earlier. The biggest cuts happened in technology jobs, with about 51,000 of those job losses in technology companies.

Some of the biggest company names have announced these cuts. Amazon said it will cut 16,000 jobs and use AI "agents" to do work instead. Pinterest said it is cutting 15% of its workers to focus on AI. Dow, a chemical company, is cutting about 4,500 jobs as it uses more AI. HP, which makes computers, said it will cut between 4,000 and 6,000 jobs and save $1 billion by the end of 2028. Other companies like Workday, CrowdStrike, and Chegg have also announced AI-related cuts.

Hiring Slowdown Signals Big Change

Companies are also slowing down how many new workers they hire. The head of OpenAI said his company will hire much more slowly because AI can do more work with fewer people. This is a big signal that companies are changing how they think about workers and money. Young workers between 20 and 24 years old are feeling this the most, with unemployment for this group hitting 9.2% in late 2024.

AI Agents Are Growing Fast

At the same time companies are cutting jobs, they are buying many more AI agents. According to a big report from Salesforce, companies plan to use 67% more AI agents by 2027. Right now, companies use about 12 AI agents on average. Almost 83% of organizations say that most or all of their teams already use AI agents. These AI agents are no longer just experiments—they are becoming the main way companies do real work.

The Connection Problem

But there is a big problem that companies are running into. When you have many AI agents working in one company, they need to talk to each other and share information. Right now, 50% of AI agents operate by themselves and are not connected to other agents. This causes problems like repetitive work and confusion.

IT leaders—the technology workers who run company computer systems—say this is very important to fix. 96% of IT leaders say that AI agents need to be connected across all systems in their company. 94% of IT leaders believe companies need to change their computer systems to be more API-driven, which means building them in a way where different parts can easily share information.

Workers Are Scared and Unsupported

Workers are having strong feelings about all these changes. More than one-quarter of employees say they do not trust their company to handle AI fairly. Many workers are worried they will lose their jobs. Studies show that fears about AI taking jobs almost doubled in 2025 compared to the year before.

The problem is worse because workers are not getting enough help. In one survey, only 45% of workers said they feel supported by their company in learning new AI skills. However, 70% of company leaders think their workers are already learning these skills and doing well. This big difference shows that companies and workers are not talking enough about training and support.

New Jobs and New Skills

Experts say that the future of work will be different, not gone. Companies will need new types of workers. New job titles are already appearing, like "agent engineers" and "escalation specialists" who help manage AI agents. These are people who understand how AI agents work and can fix problems when they happen.

The shape of the workforce might change too. In some types of work, there might be fewer middle-level jobs, but more jobs at the top and bottom levels. In other types of work, like customer service, companies might have fewer entry-level jobs but need more mid-level workers to manage the AI agents.

Rules and Responsibility

Governments are also getting involved. New regulations are being created to make sure companies use AI fairly and safely, especially in hiring and worker management. Companies now need to explain how their AI systems make decisions and need human people to check the work. This means company leaders and workers need to understand how AI works, not just use it.

What Comes Next

2026 is shaping up to be the year when AI agents stop being new and exciting experiments and become the normal way companies do business. The real test will be whether companies can make this change while also helping workers learn new skills and protecting jobs that matter.