A dot is OpenAI's always-on personal agent. You message it in ChatGPT or Slack or call it from ChatGPT, it works on its own cloud computer and browser, and it comes back with results and the decisions that need you. For one person's work, it is very good. Boring runs the work a company answers for: agents owned by a team, started where work arrives, held to the exact actions they were granted, and reviewed by the person assigned to decide. Use both. This page shows where the line sits.
Updated
If the work is yours, get a dot. You can talk to it in ChatGPT, in Slack, or on a call. It works in its own cloud browser, signing in to websites through a private form you fill in. It researches your connected apps read-only and suggests next steps. And OpenAI surrounds it with real controls, from automatic action review to app permissions that separate reading from sending. Boring answers a different question: who operates a workflow when the company has to answer for it. Its agents belong to a workspace, start from webhooks, inbound email, forms, and app events as well as schedules, call only the exact actions they were granted, and wait for an assigned reviewer who can fix the proposed call before it runs. Let your dot handle the conversation. Run what the company answers for through Boring.
| Capability | Boring AI | OpenAI's dots |
|---|---|---|
| // shape of the product | ||
| What it is | A console where a team builds, operates, and reviews automation agents. | A personal agent in ChatGPT, powered by GPT-6 Astra, with its own cloud computer and browser. |
| Who owns and directs it | A workspace: roles, private or workspace visibility, live co-editing, and versioned definitions where every change is a diff you can roll back. | A personal agent attached to its owner's account. OpenAI's messaging guide says it responds to you in Slack by default; you can instruct it to engage with others. |
| Where you talk to it | The console and workspace assistant (⌘J), plus named Slack requests, supported review/input forms, and same-agent thread follow-ups. Linked identities and agent opt-in apply. Native mobile approvals are built; store release is pending. | ChatGPT on desktop, web, and mobile, voice calls you start, and Slack. Texting via iMessage or RCS is rolling out in beta to US Pro users; approval requests link back to ChatGPT. The admin guide still describes Microsoft Teams as an invite-only alpha. |
| A computer of its own | No computer use or signed-in browsing. Agents act through connected tools and can load public web pages. | Its own cloud computer and browser, a private sign-in form for websites, Take over and Return control, and one connected personal computer if you allow it. |
| // judgment and control | ||
| What waits for a person | New agents start with every connected action waiting for approval. Owners loosen it to writes only or auto-approve eligible actions they trust, such as reads; an action Boring can't classify counts as a write. | An automatic review checks actions that could affect your accounts or share information, then lets them proceed, asks you, or hands the step to you. |
| Who decides, and what they can change | An assigned reviewer in a shared Tasks inbox can approve, modify the proposed call, or reject it, and the agent's owner or an admin can reassign it. Reviews can carry due dates, reminders, and a configured overdue outcome. | The owner opens the request and responds. Editing the proposed action first, or routing it to another reviewer, is not specified in OpenAI's public docs. |
| Hard limits on actions | Exact-action grants are enforced in code: an agent without the send action cannot send, whatever its instructions say. | App permissions can allow reading email without sending it, and admins can limit a connection to read-only or an approved set of actions. |
| Rules and procedures | Numbered instructions run as an enforced plan, one step at a time, and a step can require evidence: a saved artifact or a successful tool receipt. | Custom rules offer four choices, from "Take action without asking" to "Hand off to you". OpenAI: they "are instructions your dot tries to follow, and it can make mistakes." |
| // starting, recovering, and records | ||
| How work starts | Schedules, webhooks, inbound email, public forms, SaaS app events, the Chrome extension, the workspace API, MCP clients, and other agents. | Conversation, saved schedules, event monitoring where a connected service supports it, and MCP Events. Beyond MCP Events, a webhook URL, public form, API, or inbound email address that starts a dot is not specified in OpenAI's public docs. |
| When something fails | Resume eligible failed runs from the last completed step; in a numbered step plan, when an action's outcome is uncertain, Boring blocks automatic repetition instead of guessing. Failure reasons are classified, with the provider's reference ID. | Pause, Resume, and stop controls; stopping doesn't undo completed actions. Retries or resuming from a failed step are not specified in OpenAI's public docs. |
| Record of the work | One shared trace per run: reasoning, tool calls, approvals, and cost. Tool payloads are kept 90 days; the outcome, output, and step list stay. Run summaries export as CSV or JSON. | Activity in the desktop app shows progress, files, results, and pending requests. Enterprise admins use supported Compliance API records; OpenAI advises validating the records your workflow needs before relying on audit coverage. |
| The model | Managed Claude, GPT, Gemini, and Grok with no key, or your own Anthropic, OpenAI, Google, OpenRouter, or compatible-endpoint key. | GPT-6 Astra. OpenAI's admin guide says enterprise model controls and defaults do not apply to the main dot; delegated Work and Codex tasks follow those products' controls. |
| Price and access | Early access, on design-partner terms, with white glove if you want the agents built and operated with you. | Comes with eligible plans, rolling out gradually: Pro from $100 a month as of October 2026 (adults outside the EEA, UK, and Switzerland), Business Premium, and Enterprise once an admin turns it on. |
Details about dots come from OpenAI's own documentation at learn.chatgpt.com, checked 3 November 2026, including its October 26–30 texting announcement. Dots and individual channels are rolling out gradually. The messaging and admin guides differ on who can direct a dot in Slack; we describe the messaging guide's default without treating it as team ownership. Where public docs don't specify something, we say so. If anything here is out of date, tell us and we'll fix it.
OpenAI describes a dot as a personal agent in ChatGPT, with its own cloud computer and browser. It keeps notes on your preferences and decisions, carries context across connected channels, and checks that you've allowed it before sharing private information. Its messaging guide now says it responds to you in Slack by default and can engage with others when instructed. That expands how you work with your personal dot; team-owned tasks are a separate product.
OpenAI also offers separate Team Tasks: team-owned scheduled or event-triggered work using a team service account and configured connections, with model selection and admin audit records. If that fits your company workflow, evaluate it seriously. This page compares the personal dot with Boring, rather than claiming OpenAI has no team automation.
A company workflow has a different shape. The refund queue, vendor onboarding, and the Monday customer report each have several people who edit them, review what they did, and answer for the result, and the person who should approve a step is often not the person who set the work up. Boring puts agents in a workspace, with roles, private or workspace visibility, live co-editing, and one shared record per run. Every change to a definition is a versioned diff you can roll back.
Dots check before consequential actions, and the design is careful. An automatic review compares each action that could affect your accounts or share information against your instructions, permissions, custom rules, and built-in safety requirements, then lets it proceed, asks for your approval, or hands the step to you. Some steps you must do yourself; OpenAI's example is changing a password. When a request is waiting, you open it in Activity and respond.
Boring starts stricter and routes the decision. New agents start with every connected action waiting for approval; owners loosen that to writes only, or auto-approve eligible actions they trust, such as reads. Each request leads with the exact tool call and its scrubbed arguments and goes to the reviewer you assign. It can carry a due date with reminders, and an overdue outcome you configure: reject or skip. The reviewer can approve, reject, or modify the exact arguments before the call runs. The agent’s owner or a workspace admin can reassign the request. OpenAI's public docs don't specify editing a dot's proposed action before approving it, or routing a request to anyone other than the owner.
OpenAI gives dots real hard boundaries. App permissions can allow a dot to read email without allowing it to send, and workspace admins can limit a connection to read-only actions or an approved set. On top of those sit custom rules, with four choices from "Take action without asking" to "Hand off to you". OpenAI describes them candidly: "They are instructions your dot tries to follow, and it can make mistakes."
Boring moves more of the policy into code. Exact-action grants decide what an agent can call at all, so a send action can be withheld entirely rather than prompted against. Numbered instructions run as an enforced plan: one step at a time, no skipping ahead, each step marked complete. A step can require evidence before the plan moves on, such as a saved artifact or a successful tool receipt for an exact target. The model still uses judgment inside a step; which actions it may call, and the order of the steps, are not left to it.
A dot starts from conversation, from saved schedules, and from event monitoring when a connected service supports it, including MCP Events from servers that publish them. That covers a great deal of personal work. A company's work also arrives as a webhook from another system, an email to a shared address, a form a customer submits, or an event in a SaaS app. Boring starts runs from all of those, plus schedules, the Chrome extension, the workspace API, MCP clients, and other agents. Beyond MCP Events, OpenAI's public docs don't specify a webhook URL, public form, API, or inbound email address that starts a dot.
When a Boring run breaks, its saved plan and progress stay, and eligible failed runs resume from the last completed step. In a numbered step plan, when a write's outcome is uncertain, Boring blocks automatic repetition instead of guessing. Provider throttling and outages are waited out within a bounded window; if the provider stays down, the run ends with its progress and a clear reason. Failure reasons are classified (billing, rate limit, context size, safety stop), with the provider's reference ID. A dot can be paused, resumed, and stopped, and OpenAI notes that stopping doesn't undo completed actions; retries and recovery from a failed step are not specified in its public docs.
Sources:Meet dotsControl your dotMessage your dotText your dot — October 26–30 updateTasks and memoryComputers and appsDots admin guideApps and connectors (admin)Local access and auditTeam TasksMCP EventsPricing
Yes, and that's the split we'd suggest. Let your dot handle the conversation: your inbox, your research, your projects, the decisions only you make. Run what the company answers for through Boring: the queues, hand-offs, and recurring jobs that several people own, that start from other systems, and that need an assigned reviewer and a shared record. The two don't need to be connected to be useful side by side.
For your own work, it often is. A dot's automatic review decides whether an action can proceed, needs your approval, or is a step you must do yourself, and OpenAI layers app permissions and admin action limits on top. The difference is who gets asked and what they can do. In Boring, the request goes to the reviewer you assign, can carry a due date and reminders, leads with the exact tool call, and can be approved, rejected, reassigned, or modified before it runs. And an action an agent was never granted can't be called at all.
Yes. OpenAI's separate Team Tasks are team-owned scheduled and event-triggered tasks that run in the cloud on a team service account, with configured connections, a model choice per task and Compliance API records for admins. If your company already runs ChatGPT Enterprise, evaluate them on their own terms. This page compares the personal dot; Team Tasks merit their own evaluation.
A dot runs on GPT-6 Astra. OpenAI's admin guide says enterprise model controls and defaults do not apply to the main dot; delegated Work and Codex tasks follow those products' controls. Boring offers managed Claude, GPT, Gemini, and Grok with no key, and accepts your own Anthropic, OpenAI, Google, OpenRouter, or compatible-endpoint key. Set a workspace default, or pin a model per agent.
You can work with enabled Boring agents in Slack, but Boring does not offer voice calls. Link your Boring account, then address a published agent by its exact name in a DM or @mention. Results and optional progress return to the originating thread, where you can answer supported forms, review actions, and follow up with the same agent after the run finishes. Follow-ups use bounded run summaries rather than unrestricted chat history. The console remains the place to manage agents and review proposed definition changes before publishing.
We've tried to make it so. The official documentation was rechecked on 3 November 2026, including texting and the separate Team Tasks product. We flag the differing Slack guidance rather than repeat an absolute owner-only claim. Where the docs don't specify something, we say so. A dot wins on conversation, a computer and browser of its own, read-only proactive research, and inclusion in eligible ChatGPT plans. Dots are rolling out and changing quickly; if we got something wrong, tell us and we'll fix it.
Tell us the workflow you would hand off first — or ask about white glove and our team will build and run it with you.