Welcome to the forefront of conversational AI as we explore the fascinating world of AI chatbots in our dedicated blog series. Discover the latest advancements, applications, and strategies that propel the evolution of chatbot technology. From enhancing customer interactions to streamlining business processes, these articles delve into the innovative ways artificial intelligence is shaping the landscape of automated conversational agents. Whether you’re a business owner, developer, or simply intrigued by the future of interactive technology, join us on this journey to unravel the transformative power and endless possibilities of AI chatbots.
OpenAI is rolling out workspace agents in ChatGPT. Powered by Codex, the agents are built to automate complex team workflows and keep running on their own. Existing custom GPTs aren’t going anywhere for now.
OpenAI calls this an “evolution of GPTs.” Instead of just responding to individual prompts, the new agents handle multi-step tasks across teams. They run on Codex in the cloud and keep working even when the user is offline. Existing custom GPTs will stick around for now, and OpenAI says a tool to convert GPTs into workspace agents is in the works.
The big change from regular GPTs is that these agents are built for teams rather than individuals. Workspace agents can pull context from different systems, follow a team’s workflow, ask for approvals, and move tasks forward across multiple tools.
Each agent has its own workspace with access to files, code, tools, and memory. Agents can write and run code, use connected apps, retain information across sessions, and handle multi-step tasks. Teams can also put them on a schedule or hook them into Slack, where they pick up requests on their own.
Setting up a workspace agent is supposed to be quick. Users describe a recurring workflow or upload a file, similar to custom GPTs, and ChatGPT turns it into an agent, defining steps, connecting tools, adding skills, and testing the result. OpenAI also offers templates for finance, sales, and marketing.
OpenAI shared a few examples already running internally. A “Software Reviewer” checks employee software requests against approved tools and policies, creating IT tickets when needed. A “Product Feedback Router” watches Slack, support channels, and public forums, turning feedback into prioritized tickets and weekly summaries. A “Weekly Metrics Reporter” pulls data every Friday, builds charts, and shares a report with the team, and a “Lead Outreach Agent” researches incoming leads, scores them, drafts personalized follow-up emails, and updates the CRM.
OpenAI’s own sales team uses an agent that pulls details from call notes and account research, qualifies new leads, and drops follow-up email drafts into reps’ inboxes. The accounting team built one that handles parts of the monthly close, from journal entries to balance sheet reconciliations and variance analyses.
Knowledge inside companies is usually spread across people and systems. With workspace agents, OpenAI aims to package that knowledge into reusable workflows that follow the right process and use the right tools.
Because agents have memory and can be corrected mid-conversation, they should get better with use. Teams can build an agent once, refine it over time, and then share or duplicate it for new workflows.
Currently, teams can interact with agents in ChatGPT and Slack, with more integrations on the way. Users decide which tools and data each agent can access and when it needs permission. Sensitive actions like sending an email or creating a calendar entry can require approval, and analytics show how often a shared agent gets used.
Enterprise and Edu admins get role-based controls to set which user groups can create and share agents and which tools they can tap into.
Built-in safeguards are designed to block prompt injection attacks. Agents can be monitored through the Compliance API and paused if something goes wrong. More details are in the OpenAI Academy and Help Center.
OpenAI is launching workspace agents as a Research Preview for ChatGPT Business, Enterprise, Edu, and Teachers plans. On Enterprise and Edu plans, admins can turn the feature on through role-based controls. It’s free until May 6, 2026, after which credit-based billing kicks in.
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive “AI Radar” frontier report six times a year, full archive access, and access to our comment section.
Stay in the loop on AI. Clear, useful, no fluff.
Follow The Decoder for AI news, background stories and expert analyses.
Stay in the loop on AI. Clear, useful, no fluff.