OpenAI dots: what they are and how SMBs should prepare for always-on agents

OpenAI dots are artificial intelligence agents designed to work continuously on behalf of a person or organization. According to OpenAI's official announcement dated September 29, 2026, each dot is powered by GPT-6 Astra, has its own cloud computer and browser, learns from feedback and can connect to more than 4,000 apps through plugins.

For a small or midsize business, the shift is from one-off prompting to an ongoing working relationship. A dot can track activity, move projects forward, notice pending work and ask for a decision. That makes AI feel closer to a digital teammate, but it does not automatically make a disorganized process reliable.

What are OpenAI dots in practical terms?

A dot combines conversation, working context, connected tools and execution. It can accept a broader goal, split the work into steps, use its own computing environment and continue while the user handles other priorities.

OpenAI describes examples such as investigating bugs found in Slack, revising launch materials when scope changes, updating sales proposals and turning transcripts into content. In one early test, a dot noticed a forgotten invoice, prepared it and sent it after the user approved the action.

The core value is continuity across tasks, channels and decisions. Dots can be reached through ChatGPT, Slack and Teams and can report progress without forcing the user to restart context in a new conversation.

How is a dot different from a chatbot?

A chatbot usually reacts to a message. A dot can operate in the background, use tools, manage several workstreams and return with progress, questions or deliverables for review.

• Chatbot: answers a specific request.
• Traditional automation: follows predictable rules after an event.
• Dot: works toward a goal and chooses next steps within defined boundaries.

These models complement one another. Deterministic rules remain better for calculations, reconciliation and integrations that cannot tolerate interpretation. A dot can coordinate work, assess exceptions and prepare decisions. The article on AI agents in sales shows this distinction in a commercial workflow.

What can dots do for an SMB?

Keep sales work moving

A dot can gather account history, customer requirements and product documentation, identify gaps and prepare a proposal update. Pricing, contractual terms and customer commitments should remain subject to approval.

Organize support and operations

With controlled access, it can summarize incidents, find recurring patterns, update a work queue and flag cases for human intervention. Support needs explicit handoff rules, as explained in the guide to moving an AI conversation to a person.

Maintain content and projects

A dot can watch a calendar, turn raw material into drafts and revise assets when a decision changes. The benefit is continuity, not publishing without editorial review.

Monitor administrative gaps

It can identify missing documents, prepare payment reminders, notify owners and gather closing evidence. Sending messages, moving money or making commitments requires explicit limits.

How do security, permissions and approvals work?

Each dot uses its own cloud computer. The user's computer remains separate unless the user chooses to connect it. Organizations choose which apps are available and manage permissions through ChatGPT controls.

OpenAI says proactive research uses read-only tools when the user is not actively working with the dot. In that mode, tools cannot send messages, modify content or control a browser or computer. Actions affecting accounts or sharing information may go through auto-review and require approval. Sensitive tasks such as changing a password remain with the user.

Dots can still make mistakes. Work involving customers, contracts, payments, personal data or production systems requires review proportional to its impact.

What should an SMB define before connecting a dot?

1. Responsibility: choose a specific process rather than “run the company.”
2. Authoritative sources: identify trusted data, policies and templates.
3. Minimum access: grant only the apps and permissions required.
4. Allowed actions: separate reading, drafting, editing and sending.
5. Approvals: state which decisions always need a person.
6. Quality criteria: provide examples of acceptable work.
7. Exceptions: define when to stop, ask or escalate.

This resembles sound sales automation planning: organize the process before scaling execution.

Who can access dots?

At launch, OpenAI announced a gradual rollout to Pro, Business Premium and Enterprise users in eligible markets. Enterprise access, including Edu and Healthcare, depends on workspace administrator enablement. OpenAI also previewed specialist dots with their own identity, credentials and organizational responsibilities.

Availability, allowances and plans can change. Businesses should verify the official product page when making a decision rather than assuming that a global announcement guarantees local access.

Conclusion: dots combine autonomy with governance

OpenAI dots signal an important shift from conversational AI to continuous work with tools and context. For SMBs, the opportunity is to keep sales, support, content and administrative processes moving.

The right first step is a narrow responsibility, trusted sources, limited permissions and human approval for consequential decisions. A well-directed dot can reduce coordination work; a dot connected to a confused process will simply execute that confusion with greater autonomy.