Feature Proposal: Handle Multiple Requests in the Same ChatGPT Conversation
To the OpenAI Team
October 8, 2026
Please consider a feature that lets users submit separate, independent requests while ChatGPT is still working on an earlier request, and receive each result in the same conversation as soon as it is ready.
The problem
When I ask ChatGPT to analyze documents or conduct research, the response can take some time. While waiting, I may think of another question or task that I would like to submit without losing the flow of the conversation.
Opening another chat is an option, but it can mean specifying the same materials and background again, switching between chats, and searching for the results.
How I would like it to work
While ChatGPT is processing the first request, I can add a second and a third. Each request runs independently, and the results appear as they are completed.
For example:
-
Request A: Analyze a large set of Excel files — 3 minutes.
-
Request B: Compare contracts — 8 minutes.
-
Request C: Draft an email — 20 seconds.
In this example, the results would arrive in the order C, A, then B.
These times are illustrative. Actual processing times would depend on the requests, system load, and other conditions.
Making the feature easy to use
-
Label each result with its request name so users can clearly identify what it answers.
-
Let users choose between adding information to an ongoing request and starting a new, independent request.
-
Allow requests to use the shared conversation history, settings, and relevant materials.
-
Let users select a specific result when asking a follow-up question, and cancel each request separately.
Expected benefits and suggested first step
This could reduce the need to wait for a long task before handling a short one. It could also reduce chat switching, repeated explanations, and the effort needed to find completed results.
I suggest starting with a small trial that allows two or three requests at a time, focused on tasks such as drafting text and reviewing documents. Before expanding the feature, the trial could check whether it reduces user effort compared with separate chats and whether users can reliably match results to their requests.
Additional Information
The following notes explain the proposal’s scope and what could be checked during a trial.
The main point of this proposal
Users should be able to add new, independent requests whenever they arise, even after a long task has started. They should not need to specify every task together at the beginning.
This also serves a different purpose from asking several AI models to answer the same question. The goal here is to handle different requests within one conversation.
Relationship to existing features
OpenAI’s developer tools already provide ways to run multiple processes in parallel. Other software also offers related features that run separate tasks and return results within the same chat. This proposal is about making that kind of capability easy for ChatGPT users to access. [1] [2]
The additional benefit over using separate chats has not yet been tested. In the public documentation reviewed on October 8, 2026, I could not confirm whether the full behavior described here was already available in the ChatGPT interface.
Avoiding confusion and duplicate work
Each request should use the background information and relevant materials available when it is submitted. Information from an unrelated request added later should not be mixed into it automatically.
If a request depends on an earlier result, it should wait until that result is available before starting.
For an initial trial, I suggest limiting the feature to reading, analysis, and drafting. Editing original files or taking actions such as sending emails could be left outside the trial. The number of requests that can run at once and the applicable usage limits should also be clear.
How to evaluate the benefits
I suggest comparing the same work across two approaches: using separate chats and using multiple requests within one chat.
The comparison could look at:
-
Waiting time.
-
How often users switch between chats.
-
How often users repeat background information.
-
The effort needed to find results.
-
Mistakes in matching results to requests.
-
Usage allowance consumed.
The benefits and the level of user demand have not yet been established through testing.
Main reference materials
[1] OpenAI developer documentation on running conversations in parallel.
[2] ductor documentation on separate sessions and background tasks within the same chat.
[3] ChatGPT Projects documentation on branching a conversation into a separate chat.
Documentation reviewed on October 8, 2026.