OpenAI moves Dot agent tasks into iMessage and RCS

Dot enters the messaging inbox as a task channel

OpenAI’s Dot can now be reached through iMessage and Android RCS, allowing Pro subscribers to send it instructions alongside everyday conversations. Supported inputs include text, photos and voice notes, giving users more natural ways to describe or demonstrate a task than a conventional prompt box permits.

The practical change is the interface: task delegation no longer has to remain inside ChatGPT’s app. Writers and editors can use a familiar message thread to ask Dot to handle work, while remaining in communication channels they already check for collaborators and alerts. The feature is reported as available in the US, and one report describes the messaging option as limited to Pro subscribers.

Proactive updates turn conversations into asynchronous jobs

Dot can also contact users first when a monitored task produces a relevant update. Reported updates may include text, photos and voice notes, as well as progress messages and alerts about jobs. That makes the messaging connection more than a remote prompt box: it becomes a channel for the agent to report back.

For text-focused workflows, this can reduce the need to keep an application open or repeatedly request status. A user can assign a writing or editing task in one thread and later receive its output or an alert without manually starting another exchange. The messaging format may also make structured material easier to pass to an agent, since a screenshot can accompany a written instruction and a voice note can supply context that would otherwise require typing.

That convenience creates a new operational question. Users need to distinguish Dot’s proactive messages from ordinary contacts, especially when photos, generated text or voice notes arrive inside a personal messaging history. The evidence establishes the delivery channels and input formats, but it does not specify the labeling, consent controls or export options available for those messages.

Voice, images and text expand the brief an agent can receive

The new channel accepts more than typed prompts. Users can send Dot a photograph, a screenshot or a voice note alongside text, while the agent can respond with similar media. This gives writing and editing users a compact way to supply source material and specify a desired operation.

For editorial teams, a photo of a printed page or a screenshot of a layout can accompany a request for transcription, correction or revision. A voice note can provide a spoken brief, while a text message can preserve exact terminology or instructions. The available evidence does not claim that Dot performs any particular editing operation, so the immediate significance is input flexibility rather than a documented new writing model or editing feature.

The feature also illustrates a broader product-design pattern for language tools: the agent is moving beyond a dedicated prompt interface toward ordinary communications. That places text tools closer to where users already coordinate work, but it also makes message-level design and the handling of sensitive source material more important. Anyone using the Unicode Encoder / Decoder to normalize pasted text should still verify that copied material has not introduced invisible or nonprinting characters before sending it to an agent.

Availability limits the immediate audience

The US limitation is significant for a messaging feature because iMessage and RCS are regional platform channels, and the available reports tie access to Pro subscribers. The evidence does not state a broader rollout schedule or identify additional supported markets.

Users outside the US should therefore treat the feature as unavailable rather than assume that account status alone provides access. Nor does the evidence establish separate pricing for the messaging function, or explain what happens when a Pro subscription ends while Dot still has assigned work.

For teams evaluating agent-assisted text workflows, the release is best viewed as a new access pattern with geographic constraints. A reader can check the messaging option inside the relevant ChatGPT experience, verify whether the account is eligible, and test how Dot handles a low-risk drafting or editing request before assigning sensitive material.

Tool builders should prepare for conversational source cleanup

The release implies a concrete need for tools that clean and verify text before it enters an agent workflow. Users may combine dictated briefs, copied passages and OCR-like image content, making invisible-character inspection, text normalization and format conversion useful supporting functions.

A browser-based workflow for How to See Invisible Characters in Word (Beyond Show/Hide) can help before pasted material is submitted. The How to Remove Invisible Characters From ChatGPT Output guide is relevant when generated or received text must be prepared for publication. The Unicode Encoder / Decoder provides a practical way to encode or decode text for transport, while the URL Decoder supports cleanup when encoded strings appear in copied material.

The immediate follow-up is practical: test a harmless task through the supported messaging channel, inspect the text and media exchanged, and confirm that the returned wording is suitable for editorial use. The evidence does not announce a pending deadline or additional release date, so the feature’s availability and rollout status should be checked directly rather than assumed.