OpenAI unveils Workspace Agents, a successor to custom GPTs for enterprises that can plug directly into Slack, Salesforce and more

2026-04-22 · Source: VentureBeat · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Operations & Process Management · Depth: Intermediate, medium

Summary

OpenAI has launched "Workspace Agents," a new product available to ChatGPT Business, Enterprise, Edu, and Teachers subscribers, designed to enable AI agent workers to perform tasks across various third-party enterprise applications like Slack, Google Drive, Microsoft apps, and Salesforce. These agents, powered by OpenAI's Codex, can write and run code, use connected apps, retain memory, and execute multi-step, long-running workflows, including scheduled tasks like generating recurring reports. Unlike previous session-based AI assistants, Workspace Agents operate persistently in the cloud, allowing them to complete tasks even after the user logs off. The offering emphasizes governance, with granular admin controls over agent building, publishing, tool access, and authentication methods. OpenAI is deprecating custom GPTs for organizations, requiring them to transition to Workspace Agents, which are free until May 6, 2026, before credit-based pricing begins.

Key takeaway

For CTOs and VPs of Engineering evaluating AI integration strategies, OpenAI's Workspace Agents represent a significant shift from individual productivity tools to shared, persistent organizational resources. You should assess how these Codex-backed agents, with their enhanced persistence and integration capabilities, can automate complex, multi-step business processes, while carefully configuring granular governance controls to mitigate data security and access risks.

Key insights

OpenAI's Workspace Agents enable persistent, code-executing AI workers across enterprise apps, evolving beyond session-based GPTs.

Principles

• AI agents should be persistent and autonomous.
• Integrate AI into existing workflows, not separate interfaces.
• Robust governance is critical for agent deployment.

Method

Workspace Agents leverage OpenAI's Codex as a cloud-based code-execution substrate, enabling them to run code, use tools, and maintain memory across multiple steps and scheduled workflows, rather than just responding to prompts.

In practice

• Automate weekly reporting from dynamic data sources.
• Streamline lead qualification and follow-up processes.
• Develop agents for internal IT support or accounting tasks.

Topics

• Workspace Agents
• OpenAI Codex
• Enterprise AI Agents
• Workflow Automation
• AI Governance

Best for: CTO, VP of Engineering/Data, AI Product Manager, Director of AI/ML, AI Architect, IT Professional

Related on AIssential

• 😺 LIVE NOW: Learn Workspace Agents 101 (Build, Run, & Scale) — OpenAI's new Workspace Agents offer cloud-based, action-oriented AI integration for enterprise tools, replacing Custom GPTs.
• OpenAI's Codex update lets agents build interactive enterprise workspaces via Sites and role-specific plugins — Codex's update transforms it into an agentic AI operating environment for business, enabling precise automation and interactive app creation.
• Workspace agents — Workspace agents automate repeatable, structured workflows by integrating triggers, processes, and external tools within ChatGPT.
• OpenAI's AI data agent, built by two engineers, now serves 4,000 employees — and the company says anyone can replicate it — OpenAI's internal AI data agent demonstrates the transformative power of accessible data intelligence for enterprise operations.
• OpenAI named a Leader in enterprise coding agents by Gartner — OpenAI's Codex, powered by GPT-5.5, leads enterprise AI coding agents by enabling agentic software development with robust governance and security.
• TAI #202: GPT-5.5 Moves Codex Into Real Work — Frontier AI models are evolving into integrated work systems, emphasizing tools, memory, and permissions over raw model capability.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by VentureBeat.