ChatGPT Projects have steadily moved from being a nice organisational feature to something much more useful: a working knowledge base for real tasks.
The February 2026 update is another step in that direction.
When Projects first launched, they were mainly a way to keep related chats together. That was useful, but limited.
Since then, OpenAI has been building them into something more operational:
Now Projects can pull in source material much more flexibly.
You can add links from places like Slack channels and Google Drive files, paste notes and briefs directly as text sources, and save useful ChatGPT outputs back into the Project so strong work becomes reusable context rather than a one-off response.
That matters because most business knowledge work is messy. The useful context is rarely in one clean document. It is spread across decks, notes, old chats, draft briefs, shared drives, and informal conversations. Projects are starting to reflect that reality.
The best way to think about this update is not as another AI trick. It is an improvement in knowledge management.
A good Project can now hold:
That is far more valuable than a folder of old chats. It makes Projects a stronger place to run recurring work such as:
OpenAI also improved deep research at the same time.
You can now restrict research to specific websites and connected apps, and you can review or edit the research plan before the run starts. You can also change direction partway through.
That may sound like a small interface tweak, but it is a meaningful upgrade for real-world use.
In business settings, the problem with open-ended AI research is rarely a lack of raw information. It is usually too much information, the wrong information, or information from sources you would not normally trust.
Being able to say:
makes deep research much easier to use for structured work.
That is especially relevant for:
If your organisation has been experimenting with ChatGPT Projects, this update is a good reason to revisit how you use them.
The question is no longer just "should we keep related chats together?" It is:
The teams getting the most value from AI are rarely the ones using it for isolated prompts. They are the ones building repeatable environments around recurring work.
Projects are getting closer to being that environment.
Most organisations still underuse AI because they treat it like a search box instead of a workspace.
This Projects update pushes in the right direction. It makes ChatGPT more useful not by making it more theatrical, but by making it easier to ground work in real context and improve it over time.
That is the kind of product improvement that actually changes behaviour.
Where to next
If this topic is relevant to your team, these pages are the best next stop.
Explore how we train editorial, communications, and creative teams to use AI without flattening quality.
See how we help brand, retail, and customer-facing teams move faster without losing control of voice or quality.
See a real client example of bespoke AI training inside a fast-moving consumer business.
We tested all eight ChatGPT personalities so you don't have to. Default wins for most tasks, Candid cuts waffle by 16%, and Cynical is the sharpest strategic voice. The other five? Skip them.
OpenAI's shared Projects update makes ChatGPT much more usable for real collaboration. Teams can now share context, files, and instructions together instead of passing around disconnected chats.