Big tech companies and startups are racing to build AI agents – computer programs that can think and do things on their own without being told exactly what to do. This week, we saw major announcements and funding rounds that show how important AI agents have become.

One of the biggest news stories was that OpenAI, Anthropic, and Block teamed up to create the Agentic AI Foundation, a group that will help different AI agents work together smoothly. Think of it like making sure different car makers use the same road rules. Google also released Gemini Deep Research, a powerful tool that can read lots of information and write detailed reports for you.

On the money side, startups building AI agent tools raised millions of dollars to grow their businesses. Worktrace AI, founded by Angela Jiang who helped create ChatGPT, raised $9.3 million to help companies automate their work. Equixly, a company in Europe, raised €10 million to protect computer systems from hackers using AI agents that act like human hackers. A startup called OnCorps AI raised $55 million to help manage money using AI agents. Prime Security got $20 million to build security tools powered by AI agents.

But not everything is moving fast. Companies are still figuring out how to safely use AI agents in their work. Half of the businesses that experts talked to were still just testing out AI agents, not actually using them for real work yet. The biggest worry is security and making sure AI agents don't do something harmful. Many companies are asking whether they should build their own AI agents or use ones made by big tech companies like Microsoft or Salesforce.

Overall, this week showed that the world is moving toward AI agents as the next big thing in technology. The industry is building the foundation for AI agents to work everywhere – but companies need time to understand how to use them safely.