Enterprise AI agents go mainstream

Artificial intelligence agents are moving into workplace software, with OpenAI and Google both rolling out tools this week for companies to build and govern bots that do jobs across business systems. (openai.com) (cloud.google.com) An AI agent is a bot that does multi-step work after a prompt, instead of only answering a question once. OpenAI said on April 22 that its new workspace agents in ChatGPT can prepare reports, write code, respond to messages, and keep running in the cloud after a user signs off. (openai.com) OpenAI said the agents are available for ChatGPT Business and Enterprise workspaces and can be shared across a company in ChatGPT or Slack. The company said they work with tools including Slack, Google Drive, and Microsoft SharePoint, with admin controls and permissions set by the organization. (openai.com 1) (openai.com 2) (help.openai.com) Google made a parallel push at Cloud Next ’26 on April 22, introducing Gemini Enterprise Agent Platform as a system for building, scaling, governing, and optimizing agents. Google said the platform combines Vertex AI with new agent integration, orchestration, DevOps, and security features. (cloud.google.com) (blog.google) Google framed that launch as part of what it called the “Agentic Enterprise,” with customer examples including Capcom and Citi using agents for tasks such as game testing and financial-advice workflows. The company used its annual cloud conference to position agents as a core part of enterprise software, not a side experiment. (cloud.google.com) (blog.google) The new selling point is control. OpenAI described workspace agents as an evolution of GPTs with organization-level permissions, while Google emphasized security, DevOps, and governance alongside model access. (openai.com) (cloud.google.com) Other vendors are packaging the same idea as managed software for companies that want AI to act inside existing workflows. Faculty said on April 23 that it launched a Frontier Plug-in Marketplace so customers can add modular plug-ins inside current deployments without refactoring systems. (prnewswire.com) (faculty.ai) Strider has been taking a similar path in strategic intelligence, where companies use AI systems to sift public data for geopolitical and supply-chain risk. The company said in May 2025 that new capabilities for its Spark product would turn its global intelligence platform into faster, actionable outputs for customers. (prnewswire.com) (striderintel.com) The pattern across those launches is that vendors are no longer selling only chat windows. They are selling shared bots, plug-ins, and admin policies meant to fit procurement, compliance, and day-to-day work inside large organizations. (openai.com) (cloud.google.com) (prnewswire.com) The next test is whether companies trust these agents with more than low-risk tasks. This week’s launches show the market is now being pitched less as chatbot access and more as governed software that can take actions on a team’s behalf. (openai.com) (blog.google)