Dot 4 / Boundaries

Set boundaries before giving more responsibility

Begin with comfortable preferences, then explicitly choose what dot may do as the work develops.

Start with a preference you can say out loud

“Show me drafts before sending them” is a reasonable way to begin. So is “You can update this private checklist from the notes I share.” You do not have to anticipate every possible action before you start a conversation.

These are beginner preferences you can explicitly revise. As the work becomes clearer, choose a scope for any action or recurring responsibility you want to authorize. Dot’s actual access, approval checks, and safeguards still apply.

When you need detail, name three kinds of action

A useful brief answers three different questions: what may proceed, what needs approval, and what must not happen. Avoid a broad “do whatever is needed” instruction. It hides the decisions you should make deliberately.

For a follow-up assistant, the arrangement could be:

  • May proceed: read the selected, authorized records and prepare a private summary.
  • Ask first: send a message, make a purchase, delete a record, change access, or change a production system.
  • Do not do: contact a customer about a legal dispute or include unrelated personal information.

Within this example, the third list wins if it conflicts with the other two. A prohibition does not become an approval request just because another sentence sounds permissive. If you want to change a boundary, say which instruction you are replacing and what the new scope is.

Be specific about approval

“Ask before important things” leaves “important” undefined. Name the action, recipient or target, scope, and any limit. For a proposed message, review the actual recipient and wording. For a proposed edit, review the specific file and change.

Approval for one action should not be silently stretched to cover unrelated future work. A specific ongoing instruction can cover future actions within its stated scope; work outside it needs another decision. Our builder starts with conservative defaults and flags ambiguous permission text rather than broadening it.

Review actual product controls separately

A brief expresses your intentions. It does not enforce permissions, connect accounts, or override product safeguards.

OpenAI documents separate plugin permissions, optional Custom Rules, and built-in safety checks. Conversation is a place to express preferences and scoped instructions; it is not proof that settings were changed. Custom Rules do not grant app access or disable mandatory safeguards. Official Dots controls guide.

OpenAI also distinguishes proactive research from authorized action: research tools cannot directly message people, modify connected content, or control a browser or computer. Official privacy FAQ.

Check the current settings and documentation in your account. Do not assume a pasted paragraph has configured them.

Use conversation for tone and update preferences. To save an action rule, OpenAI currently documents Settings → Personalization → Permissions → Custom rules. These optional rules can allow a specified action, require your instruction or approval, or hand it to you. They do not grant app or computer access, and required safeguards still apply. Workspace policy can disable Custom Rules; the rules themselves can also be misapplied. Review important actions instead of treating a rule as a guarantee.

For important actions, ask dot to show the result or confirmation, and check it before relying on it.

Decide when work should stop

Write interruption rules for missing authorization, conflicting facts, a request outside scope, an unexpected cost, or an action that cannot be undone. Give the helper a useful alternative: explain what is blocked, show the evidence, and suggest a smaller permitted next step.

“Stop and ask” is easier to act on when it includes what the question should contain.

To stop existing work, check its actual controls. Pause stops the main task; delegated tasks in Activity and recurring tasks in Scheduled need separate attention. Confirm relevant monitors or automations have stopped. Completed actions are not automatically undone. Practical stopping guidance.

Try this now

Take one responsibility and list three actions it might lead to. Assign each to “may proceed,” “approval,” or “prohibited.” If you cannot decide, keep it out of autonomous work for the first test.

Read the resulting brief for contradictions. Pay attention to phrases such as “always,” “anything,” and “without asking.”

Common mistakes

Confusing instructions with enforcement. Review settings as well as wording, then inspect consequential work.

Allowing a whole category after one example. Keep permission narrow enough that you can describe what you approved.

Expecting perfect compliance. Clear boundaries help communicate intent; they are not a guarantee. Keep first tests reversible and reviewable.

Your next dot

A working arrangement needs a review loop. Continue to Give feedback that improves the next result.

Official references

Source check: October 14, 2026. A source check is not hands-on testing.