2025 became a milestone year for AI agents in business, which are smart computer programs that can work independently to help companies complete important tasks. Unlike basic chatbots that only answer questions, AI agents can make decisions, create plans, and take actions on their own while humans watch over them. This weekly update covers the biggest news about how businesses are using AI agents to change the way they work.

Major Company Launches Show Progress in AI Agents

Zoom, the popular video meeting company, launched a powerful new version of its AI tool called AI Companion 3.0. This new version goes far beyond summarizing meetings. It can now browse the internet, help with writing, and handle complicated workflows that involve multiple steps. Zoom's AI Companion also works with other apps and tools that businesses already use. This shows that AI agents are becoming more useful and connected to the real tools companies depend on every day.

Capital One, a major credit card company in the United States, created an AI agent called Chat Concierge to help car dealerships. The company designed this agent specifically for customers looking to buy cars. The agent can answer questions about different vehicles and help set up appointments with salespeople. The results were impressive—the agent increased customer engagement by a lot and helped convert 55% more leads into actual buyers. However, Capital One had to work hard after the launch to fix problems like slow response times. The company improved the system five times over, which shows that even successful AI agents need regular improvements.

IBM Releases Open-Source AI Agent

International Business Machines Corporation (IBM), a major technology company, released an open-source agent called CUGA that other companies can use and modify. CUGA is designed to handle complex tasks by working with APIs and generating code. In tests, CUGA succeeded in completing 61.7% of tasks in one test and 48.2% in another. While these numbers might seem low, experts say they could be acceptable for certain types of work. The release shows that the AI agent field is growing and that companies are sharing their technology with others.

Major Business Platforms Leading the Way

Salesforce, a giant cloud software company, has become one of the biggest leaders in AI agents with its Agentforce platform. Since launching in October 2024, Salesforce closed 18,000 Agentforce deals, which means 18,000 companies bought or signed up to use the platform. The platform helps businesses in customer service, marketing, sales, and other areas. Even more important, the United States government started using Salesforce's Agentforce. The IRS, which is America's tax agency, began deploying the platform across multiple divisions, marking the government's first widespread use of AI agents. This shows that AI agents are trusted enough for even sensitive government work.

Challenges and Realistic Expectations

While the news about AI agents sounds exciting, the reality is more complex. Research from the consulting company Deloitte found that while 30% of organizations are exploring AI agents, only 11% are actually using them in production—meaning real work, not just tests. Many companies are learning that building the agent is only the first step. After launching, they must monitor performance, fix problems, and continuously improve the system. One consulting firm found that 88% of large companies use AI in at least one area, but only 20% have AI tools that work across different departments. The technology is complex, and integration takes time and money.

Future Growth and Opportunities

Experts are optimistic about the future. Gartner, a major research firm, predicts that by 2028, AI agents could handle 15% of daily business decisions. Google Cloud reports that one-third of all business software applications are expected to include agentic AI features by 2028. A survey from Zapier found that enterprise investment in AI agents is expected to increase sharply in the coming year. This means more companies will spend money on AI agents. Looking ahead, businesses are moving from just experimenting with agents to actually using them at scale to solve real business problems. The companies achieving the best results are those that start small with low-risk projects, measure the outcomes carefully, and then expand to more complex uses.