ChatGPT Dots for Business

Deploying ChatGPT dots for business makes sense as a single-seat pilot on a Business Premium account for bounded, read-heavy work with approval rules on. OpenAI announced dots on September 29, 2026, as persistent, always-on agents powered by GPT-6 Astra that operate within dedicated cloud environments. Because these agents run continuous background tasks, companies need strict approval controls before granting broad access.

Access depends directly on your workspace tier. Business Premium seats receive access across all supported ChatGPT regions, while Enterprise workspace administrators must manually enable the beta feature. Pro subscriptions also include dots, but OpenAI's Pro rollout excludes the EEA, Switzerland, and the UK.

A broad rollout is premature while OpenAI's usage limits for dots are unpublished. Plans include an allowance for deeper work, with extended limits for the first month after launch, but OpenAI has not published the usage terms that follow or the price of additional dots.

Who Can Turn Dots On Today

Yes, if one person can sit on a Business Premium seat: pilot one dot on a bounded, read-heavy job, such as watching customer feedback or updating a proposal, with approval rules on for anything it sends or buys.

OpenAI announced dots on September 29, 2026, and its getting-started article for dots says they are rolling out gradually to Pro and Business Premium users. For ChatGPT Business accounts, Premium seats include access to your first dot across all supported geographic regions. In contrast, OpenAI excludes Pro subscribers located in the European Economic Area, Switzerland, and the United Kingdom from individual access.

Enterprise, Edu, and Healthcare workspaces can evaluate dots through an opt-in beta program. Workspace administrators must manually enable this feature, as OpenAI leaves it turned off by default across all institutional accounts. Access is expanding in waves, meaning eligible accounts may wait several days before the feature appears.

Setting up a dot requires the ChatGPT desktop application on macOS or Windows, or the desktop web interface. You cannot create a dot on mobile, and dots are not supported on mobile web. Once configured on desktop, users can interact with their dot through the ChatGPT mobile app or connect messaging tools like Slack and Microsoft Teams.

Dots are unrelated to the dot indicators in the ChatGPT interface.

Who this is not for: Organizations that require unattended autonomous task execution without human oversight should not adopt dots today. Teams needing fixed, predictable monthly costs across a whole team should also pause, as OpenAI has not published usage limits beyond the initial rollout period. Those organizations should continue using standard ChatGPT Business seats with human-directed prompts.

What would change our verdict: Broader team rollouts become advisable if OpenAI publishes clear post-launch consumption caps, defines pricing for adding multiple dots per seat, and opens administrative specialist dot controls to all commercial workspaces.

• Business Premium: included at no extra cost in all supported ChatGPT regions
• Enterprise, Edu, and Healthcare: available as an opt-in beta that starts turned off
• Pro: included for individuals, excluding the European Economic Area, Switzerland, and the United Kingdom
• Free, Go, and Plus: dots are not listed as available on these tiers

ChatGPT Dots for Business Pilot Costs

Running a business pilot for ChatGPT dots requires a minimum commitment of two seats on ChatGPT Business, with at least one user assigned to a Premium seat.

OpenAI structures ChatGPT Business seats around Standard and Premium tiers. Standard seats cost $25 per user per month on monthly billing, or $20 per user per month billed annually. OpenAI does not list dots for Standard seats. Premium seats cost $125 per user per month on monthly billing, or $100 per user per month billed annually. Premium seats include 5x more usage than Standard seats, remove the 5-hour usage limit, and include your first dot at no extra charge.

Seats can be mixed and reassigned, so a pilot can put one person on a Premium seat and keep the second required user on a Standard seat. Individual accounts can also access dots through ChatGPT Pro tiers, which cost $100 per month for Pro 100, $200 per month for Pro 200, and $500 per month for Pro 500. Pro subscriptions are billed monthly only and offer no annual billing discount.

Astra Ultrafast is included only on Pro 500. Among individual plans, the ChatGPT pricing page lists a dot only on the Pro plan card, not on Free, Go ($8 monthly), or Plus ($20 monthly). Enterprise workspace pricing is unpublished and requires direct discussion with OpenAI sales.

Subscription rates and usage policies change over time, so verify current figures directly on the official ChatGPT pricing page before making purchasing decisions.

For the seat math against a personal plan, see whether to choose ChatGPT Pro or Business.

• ChatGPT Business Premium: $125 per user per month, or $100 per user per month billed annually; includes your first dot
• ChatGPT Business Standard: $25 per user per month, or $20 per user per month billed annually; a dot is not listed
• ChatGPT Pro 100: $100 per month on monthly billing, includes your first dot
• ChatGPT Pro 200: $200 per month on monthly billing with higher usage allowances
• ChatGPT Pro 500: $500 per month on monthly billing, includes Astra Ultrafast

Recommended Workflows for Initial Testing

Initial business pilots should restrict dots to bounded, read-heavy workflows where human operators inspect all deliverables before execution.

Each dot operates on its own cloud computer with its own cloud browser, powered by GPT-6 Astra. The agent connects to over 4,000 applications through the OpenAI plugin ecosystem and works toward defined goals around the clock. Because dots maintain context across desktop, web, Slack, and Microsoft Teams, operators can assign background research projects without monitoring active chat windows.

OpenAI highlighted several practical examples in its introducing dots announcement. A developer's dot monitors customer feedback, builds bug fixes, runs tests, and submits pull requests with attached walkthrough videos. A launch manager's dot rewrites marketing collateral when project scopes change. A sales lead's dot refreshes sales proposals and creates functional proofs of concept. A scientist's dot recalculates analytical models as new experimental data arrives.

Content teams can instruct a dot to identify video highlights, compile show notes, and draft social posts. In another example, an early tester's dot noticed he had forgotten to invoice a publication, prepared the invoice, and sent it after his approval. OpenAI explicitly warns that dots can make mistakes, advising users to review all consequential work before delivery.

• Engineering triage: monitoring customer bug reports, running automated test suites, and preparing pull requests
• Marketing updates: revising collateral, presentation decks, and launch documentation following scope revisions
• Sales operations: updating client proposals and assembling technical proofs of concept
• Administrative finance: identifying unbilled accounts and drafting invoices for human review

Governing ChatGPT Dots for Business Workflows

Securing ChatGPT dots in commercial environments requires configuring Custom Rules, enforcing auto-review checkpoints, and inspecting the Activity View regularly.

Custom Rules govern whether your dot can act independently, seek permission, or hand tasks back to human staff. Operators can choose between four explicit behaviors: Take action without asking, Take action if pre-approved, Ask before taking action, and Hand off to you. Pre-approved means you asked for that action in your prompt. These rules restrict what the agent can share, purchase, or access across external accounts.

Auto-review runs before actions that could alter external systems or share sensitive company records. The safety system evaluates planned actions against user instructions, Custom Rules, and built-in safety requirements. Critical actions, such as changing account credentials or passwords, always remain restricted to human users. When signing into supported websites, dots can use saved passwords without exposing raw credentials to the model.

Activity View shows your dot's progress, including background work, and lets you redirect it; the dot's profile lists activity under In progress, Scheduled, and Completed. You can open your dot's cloud computer at any time from desktop to observe its work directly, while mobile access opens in your direct control. Local computer access starts disabled by default. If you confirm Allow access in the ChatGPT desktop app, your dot can create Work or Codex tasks, use local skills, and use your local browser when its cloud browser is blocked. Choosing Revoke access stops it.

Data privacy policies protect commercial workspaces. OpenAI does not use content from ChatGPT Business, Enterprise, or Edu workspaces to train models by default. The system also does not train models on proactive research queries or internal notes that dots write to themselves. When dots gather information through connected apps, disconnecting the app does not erase accumulated data. To permanently delete saved memories, conversations, and scheduled tasks, users must reset the dot through its profile menu.

• Custom Rules: four distinct permission states covering what your dot can share, purchase, or access
• Auto-review: safety checks that keep credential changes and high-risk actions under human control
• Activity View: task monitoring across In progress, Scheduled, and Completed stages
• Local computer access: opt-in connection that remains disabled until explicitly granted in desktop settings
• Workspace privacy: default exclusion from model training across Business, Enterprise, and Edu workspaces

ChatGPT Dots Compared to ChatGPT Work

ChatGPT dots serve as persistent, autonomous agents that manage ongoing background projects, whereas ChatGPT Work operates as a task-oriented agent focused on delivering specific multi-step documents.

OpenAI launched ChatGPT Work on July 9, 2026, as reported by The Next Web. ChatGPT Work replaced the older ChatGPT agent, which is no longer available as noted in the agent transition article. Work is built for finished deliverables, such as a spreadsheet or a Site. It operates using its own cloud browser and pauses whenever it requires sign-in, clarification, or confirmation.

While ChatGPT Work focuses on completing discrete multi-step files, a dot acts as an ongoing collaborator that operates 24/7. Your dot can initiate projects inside ChatGPT Work or delegate technical software tasks to OpenAI Codex. When your dot starts tasks in Codex or ChatGPT Work, those operations count toward your standard account usage limits. In contrast, regular conversations with your dot do not count against ChatGPT usage caps.

• ChatGPT Dots: always-on background agent, maintains persistent cross-channel memory, delegates to Work and Codex
• ChatGPT Work: discrete multi-step deliverables, generates presentations, spreadsheets, reports, and Sites
• OpenAI Codex: cloud-based coding environment dedicated to technical development, testing, and script execution

Specialist Dots and Enterprise Integration

Specialist dots represent dedicated institutional agents designed with distinct corporate identities, provisioned credentials, and deep connections to internal systems of record.

While primary dots act on behalf of individual workers, specialist dots assume formal departmental responsibilities within an organization. OpenAI engineering teams work directly with enterprise pilot participants to define specific operational remits, approved toolkits, and human review boundaries. These specialist agents incorporate lessons from internal testing across OpenAI departments, including procurement, invoice processing, email marketing, customer support, and commercial contracting.

OpenAI is collaborating with Microsoft to integrate specialist dots into governance frameworks managed through Microsoft Agent 365. For broader industry context, TechCrunch reported on dots' animated avatar presentation and highlighted Meta Muse as a rival agent. VentureBeat reported that Meta Muse, Instinct, Manus 2.0, Grok Bot, and OpenClaw compete in this market. VentureBeat also noted that ChatGPT Space is replacing Library for Pro, Business, and Enterprise tiers. On the same day, OpenAI also announced GPT-6.1 Sol.

Specialist dots remain restricted to private enterprise pilots today. OpenAI has not published public onboarding timelines or commercial pricing for specialist agents. Organizations outside these private pilots must evaluate individual dots on Business Premium while waiting for broader enterprise availability.

• Dedicated identity: distinct corporate profile and provisioned access credentials for organizational accountability
• System integrations: direct data connections to internal enterprise software and records
• Microsoft Agent 365: collaborative integration with Microsoft governance and security infrastructure
• Private pilot status: access restricted to direct engineering engagements during initial rollout

Scoping a Controlled Pilot Deployment

Deploying ChatGPT dots safely requires defining narrow operational boundaries, enforcing strict approval rules, and auditing data permissions before expanding access across teams.

At Layer3Labs, we scope AI pilots inside client workflows, and we have not run a dot pilot: dots launched on September 29, 2026, so real-world deployment reports are not in yet. OpenAI has published feature descriptions and safety guidelines, but independent performance verifications across live commercial environments remain forthcoming.

Start small. Put one person on a Business Premium seat and keep the rest of the team on Standard seats. Set Custom Rules to Ask before taking action for anything your dot shares or buys. Leave local computer access off, so your dot works only on its own cloud computer.

An AI workflow audit picks the first background task worth handing to a dot. For the wider rollout, see how to govern autonomous AI agents for business and how to limit shadow AI agents.