OpenAI launched "company knowledge" enterprise search tool for ChatGPT Business, Enterprise and Edu customers
News

OpenAI Pushes Into Enterprise Search With Company Knowledge

3 minute read
Siobhan Fagan avatar
By
SAVED
OpenAI enters enterprise search with GPT-5-powered “company knowledge,” linking ChatGPT to workplace apps — but can it win big business trust?

In Brief 

  • ChatGPT moves into enterprise search with latest announcement.
  • GPT-5 provides business-specific responses with clear citations to internal knowledge assets.
  • IT admins gain granular control while users get permission-based access.

Just two days after OpenAI challenged Google Chrome with the release of ChatGPT Atlas, the company set its eyes on the enterprise search market with the introduction of  "company knowledge." Company knowledge connects ChatGPT Business, Enterprise and Education customers to workplace applications such as Slack, SharePoint and Google Drive, allowing users to receive answers based on their organization's internal information. 

According to OpenAI, the new capability is powered by a specialized version of GPT-5 trained to analyze multiple sources and provide comprehensive answers, complete with citations. The company knowledge feature respects existing permissions, ensuring users only access information they're already authorized to view.

OpenAI's Enterprise Push

Keeping up with all of OpenAI's product announcements and partnerships could be a full-time job.

The company has accelerated its business and enterprise product releases in 2025, with capabilities like role-based access control, project sharing and the introduction of connectors to common enterprise apps all signs of the company's ambition to claim part of the lucrative business market.

Its strategy appears to be succeeding. In June, the company announced it had 3 million paying business users across its Enterprise, Education and Business (then called Teams) licenses. It's unclear how many companies that figure represents, although COO Brad Lightcap told CNBC in June it was signing nine enterprise customers per week, including businesses in highly regulated sectors like financial services and health care.  

With company knowledge, OpenAI will compete with Glean, Slack and Google's Gemini Enterprise. The release taps into the current zeitgeist around "context engineering," the practice of curating and controlling the information provided to an AI model to improve the relevance of outputs.  

How OpenAI's Company Knowledge Compares to the Competition

The functionality of company knowledge — at least on paper — sounds comparable to its closest competitors.

One area where there does appear to be differentiation is that while Glean, Slack and Gemini all have a knowledge graph at their foundation, OpenAI makes no mention of a knowledge schema or graph at company knowledge's core in its announcement. If that is the case, it means company knowledge may miss the relational contexts which are one of the selling points of modern search. 

Knowledge graphs can conduct entity-aware searches, meaning a search for a person's name will acknowledge the person as an entity and not just as a name to be found in documentation. They also aid in disambiguation and relationship queries.

PlatformCore CapabilitiesSecurity & Compliance
OpenAI Company Knowledge
  • Connects to Slack, SharePoint, Google Drive, GitHub, Gmail, Outlook, HubSpot, Salesforce
  • Powered by GPT-5 with clear citations across multiple sources
  • Respects existing permissions
  • No training on customer data; industry-standard encryption
  • IP allow-listing and Enterprise Compliance API access
Glean
  • Search, Assistant, and Agents powered by proprietary knowledge graph
  • Permissions-aware results with semantic understanding
  • Agents automate multi-step workflows
  • TX-RAMP Level 2 certified
  • ISO 27001 certified
Slack Enterprise Search
  • Unified search across conversations, tools, and data
  • Integrates with Asana, Box, GitHub, Google Drive, Jira, Salesforce
  • Built-in Slack AI agents for summarization, search, and context generation
  • Encryption at rest and in transit by default
  • Enterprise Key Management (EKM), audit logs, native DLP
  • Compliant with ISO 27001, 27017, 27018, 27701, 42001, SOC 2/3, FedRAMP Moderate
Gemini Enterprise
  • Connects to Workspace, Microsoft 365, Salesforce, SAP, Box
  • Pre-built agents for research, analytics, and automation
  • Gemini 2.5 Pro and Flash models under unified governance
  • Secure-by-design Google Cloud infrastructure
  • Customer data not used for model training or advertising
  • FedRAMP and HIPAA compliant; CMEK and audit logging supported
  • Includes Model Armor for safe AI use

Beyond these technical capabilities, customers seeking a more modern enterprise search experience may also end up deciding based on a few other factors: 

  1. Model preference and ecosystem alignment
  2. Integration depth and breadth
  3. Trust and governance

On model preference and ecosystem alignment, the decision often comes down to existing infrastructure. Companies already using Google Workspace may find the move to Gemini Enterprise the natural choice, while those with Microsoft/Azure commitments might favor OpenAI.

Learning Opportunities

On integration depth and breadth, the landscape is more nuanced than a simple feature count might suggest. While Glean, Slack and Gemini boast mature connector ecosystems, they also benefit from focused product strategies. OpenAI, by contrast, positions ChatGPT as a general-purpose AI assistant where company knowledge sits alongside coding, content creation, data analysis and image generation.

But on the final factor, OpenAI is likely running at a disadvantage. Although company knowledge checks many of the functionality and security boxes, reputational exposure remains a key consideration for enterprise buyers. OpenAI’s recent consumer-facing controversies — from the Sora 2 video creator copyright dispute to its foray into adult-oriented content — may prompt some organizations to question its ability to maintain clear boundaries between experimentation and enterprise. To earn lasting trust, OpenAI will need to demonstrate that its consumer ambitions neither bleed into nor overshadow its enterprise-grade reliability.

fa-regular fa-lightbulb Have a tip to share with our editorial team? Drop us a line:

About the Author
Siobhan Fagan

Siobhan Fagan is the editor in chief of Reworked and host of the Apex Award-winning Get Reworked podcast and Reworked's TV show, Three Dots. Connect with Siobhan Fagan:

Main image: OpenAI
Featured Research