ChatGPT Connectors bring workplace files into business AI

OpenAI plans to beta test ChatGPT Connectors for select ChatGPT Team users, starting with Google Drive and Slack. The feature is meant to let ChatGPT answer workplace questions using permitted internal files, spreadsheets, presentations, and Slack conversations.

WTF Index NEUTRAL
◄ Terminator 1 Idiocracy 1 ►

A routine workplace AI integration with mild privacy and dependency implications but no clear dangerous or degrading outcome.

ChatGPT Connectors bring workplace files into business AI

OpenAI is preparing to test a new way for business users to bring workplace knowledge into ChatGPT. The feature, called ChatGPT Connectors, is planned as a beta for select ChatGPT Team users and will initially support Google Drive and Slack.

The idea is straightforward: instead of asking ChatGPT only about general information, employees could ask questions that draw on internal material their company already stores in common work tools. That could include files, presentations, spreadsheets, and Slack conversations, depending on what the company chooses to sync and what each employee is allowed to access.

What ChatGPT Connectors Is Designed To Do

According to a document viewed by TechCrunch, ChatGPT Connectors will let ChatGPT Team subscribers connect workspace Google Drive and Slack accounts to ChatGPT. Once connected, ChatGPT can respond to questions with help from the information inside those accounts.

The document describes the goal in plain terms:

“This will allow employees using ChatGPT to easily make use of internal information similar to how they can use world knowledge via web search,”

That framing matters because many business questions are not answered by the open web. They depend on internal documents, project discussions, spreadsheets, presentations, and company-specific context. ChatGPT Connectors is OpenAI’s attempt to make that material usable inside ChatGPT without requiring employees to leave the chat interface.

OpenAI also plans to expand ChatGPT Connectors to other platforms in the future, including Microsoft SharePoint and Box, according to the same document. For now, the beta is focused on Google Drive and Slack.

How The Beta Works For ChatGPT Team

The beta is launching for select ChatGPT Team users. All users in a participating ChatGPT Team workspace will get access to the model through OpenAI’s ChatGPT apps, according to the document.

The feature is powered by a version of OpenAI’s GPT-4o model that can refine answers using “internal [company] knowledge.” To do that, the model searches and “reads” internal information that may be relevant to a user’s prompt.

OpenAI creates a search index by syncing an encrypted copy of company files and conversations on ChatGPT’s servers, according to the document. When ChatGPT responds, users may also be able to inspect related information through a sources button at the bottom of an answer.

The document says:

“Additional related information which [sic] the model did not directly make use of is accessible by clicking on the sources button at the bottom of each response,”

It also says that, when appropriate, the model will respond directly with a list of relevant results. In practice, that means ChatGPT Connectors is not only positioned as a question-answering feature, but also as a way to surface internal search results inside a chat workflow.

Permissions Are A Central Part Of The Pitch

Business data access is the most sensitive part of any workplace AI tool. The document highlights that Slack and Google Drive permissions are “fully respected” and “kept continuously up to date.”

According to the document, ChatGPT Connector syncs Slack private channel memberships, Drive file permissions, and directory information. Employees will not be able to discover content through ChatGPT that they cannot access in Google Drive or Slack.

Administrators will also be able to choose which Slack channels and Google Drive files are synced. That gives companies a control layer over what becomes available to ChatGPT Connectors in the first place.

There is a practical consequence to this permission model. OpenAI said in the document that employees may receive “substantially different” answers to the same ChatGPT prompts. That can happen because two employees may not have access to the same internal files or Slack conversations.

What ChatGPT Connector Cannot Access

The beta has several technical limits. The document says images in Google Drive files are not supported. That includes images inside Google Docs, Google Slides, PDFs, Word documents, PowerPoint presentations, and plain text files.

The feature also has limits around spreadsheets. ChatGPT Connector can only “read” data in Sheets and Excel workbooks, not analyze it.

Slack access is limited as well. ChatGPT Connector cannot retrieve Slack DMs or group messages, and it will ignore messages from Slack bots.

Those limits define the early shape of the product. It can connect ChatGPT to a company’s internal knowledge base across selected tools, but it is not presented as a complete replacement for every workflow inside Google Drive or Slack.

Why This Matters For Enterprise Search

ChatGPT Connectors is part of OpenAI’s broader effort to make ChatGPT more useful inside business software stacks. The source article notes that some companies have expressed reservations about allowing ChatGPT to access sensitive business information, while others have adopted the technology more readily.

If the beta performs well, ChatGPT Connectors could make ChatGPT more valuable for everyday workplace questions. Employees could ask about internal material without manually searching through files and conversations first.

The feature could also create pressure for AI-powered enterprise search platforms like Glean. ChatGPT Connectors is moving into a similar problem space: helping employees find and use company knowledge spread across multiple workplace systems.

Companies that want to participate in the beta are being asked to provide OpenAI with 100 documents, spreadsheets, presentations, and/or Slack channel conversations. OpenAI said in the document it will not directly train on that information, but may use it “as input to synthetic data generation” that might be used in training.

The document also states:

“No data synced from Google Drive or Slack will be used for training,”

OpenAI did not respond to multiple requests for comment, according to TechCrunch.