AI Permissions & Safety

Before performing an action, ReactionBot checks whether the requesting user is authorized to use the matching feature.

It also checks whether ReactionBot has the Discord permissions required to complete the action.

A request can be understood correctly but still be rejected if either the user or the bot is missing a required permission.

ReactionBot AI is intended for server administrators and authorized moderators.

The requesting user must have the same level of authority required by the equivalent slash command or dashboard action.

Members without the required moderation permissions cannot use AI wording to access restricted ReactionBot features.

ReactionBot must have permission to view the channel where the request is made and access the channels affected by the action.

Depending on the feature, ReactionBot may also require permissions such as Manage Messages or Manage Roles.

Removing a required bot permission can cause an AI request to fail even when the requesting moderator has sufficient access.

Actions involving another member remain subject to Discord role hierarchy.

ReactionBot may be unable to manage a member whose highest role is positioned above the ReactionBot role.

The requesting moderator must also have sufficient authority over the targeted member according to Discord's role hierarchy.

Place the ReactionBot role above any member roles it may need to manage.

Moving the ReactionBot role higher does not grant additional Discord permissions by itself, so the required permissions must still be enabled.

Administrator access is not always necessary, but the bot must have every permission required by the features you use.

ReactionBot must be able to view the channel and read the information needed to apply or enforce a reaction rule.

A rule may fail or stop working if ReactionBot is denied access through channel-specific permission overrides.

Check both the server role permissions and the affected channel's permission settings when troubleshooting.

ReactionBot AI access does not automatically unlock every Free or Premium feature.

A server-wide action that normally requires Premium will still require Premium when requested through AI.

ReactionBot validates both AI access and the feature access required by the requested action.

The requesting account or server must have active ReactionBot AI access before an AI message can be processed.

Having ReactionBot Premium does not automatically provide AI access unless it is specifically included with the selected plan.

Having AI access also does not automatically grant Premium server-wide controls.

When a message contains multiple actions, ReactionBot validates each instruction separately.

One action may pass while another fails because of different permissions, plan requirements, role hierarchy, or missing details.

ReactionBot's response should identify which actions were completed and which could not be performed.

ReactionBot should not guess which member, emoji, channel, duration, or scope you intended when the request is unclear.

For example, a request such as “restrict him” may be rejected when no specific member can be identified.

This reduces the risk of applying a moderation action to the wrong user or across the wrong scope.

Server-wide changes affect more channels and therefore require clear wording and the appropriate feature access.

Use phrases such as “across the server” or “server-wide” when that is the intended scope.

Always review the confirmation carefully after requesting a server-wide change.

ReactionBot AI cannot override Discord permission restrictions, role hierarchy, channel overrides, plan limits, or unsupported feature boundaries.

Changing the wording of a request does not change whether the action is authorized.

When a request is denied, use the response to identify the missing permission, role position, access level, or required detail.

After ReactionBot processes an AI request, confirm that the correct action, emoji, member, channel, duration, and scope were used.

Multi-action requests should be reviewed in full because some actions may succeed while others fail.

Use the dashboard or equivalent slash command to correct any setting that does not match your intended configuration.

Limit powerful moderation permissions to trusted administrators and moderators.

Keep the ReactionBot role high enough to manage the intended members, but avoid granting permissions that are not needed.

Use clear prompts, review confirmations, and test major server-wide changes in a controlled channel before relying on them broadly.

Continue to Limitations to learn about supported actions, request limits, interpretation boundaries, and cases where a standard command may be more reliable.