Document360 Integration

Connect Document360 to access your knowledge base articles and categories directly in AI conversations. GPT Workbench integrates with Document360 to bring documentation content into threads as context blocks and through search tools, enabling documentation-grounded AI responses.

Overview

Document360 is a knowledge base platform used by support teams, product teams, and documentation teams to maintain structured documentation. The GPT Workbench integration provides two capabilities:

• Context blocks -- Attach specific articles or entire categories from your Document360 knowledge base to a thread. The AI receives the full article content and can reference it throughout the conversation.
• Tools -- Enable Document360 tools to let the AI search and retrieve articles on demand during a conversation.

This integration is particularly valuable for customer support teams who need AI responses grounded in official documentation, and for product teams who want to reference specifications without leaving the conversation.

Setup

Prerequisites

• Your organization administrator has enabled the Document360 integration
• You have a Document360 account with API access
• Your Document360 project has published articles

Connect Your Account

1. Go to Settings > Integrations
2. Locate the Document360 section and click Connect
3. Authenticate with your Document360 credentials
4. Select the project to connect (if you have multiple projects)
5. A green checkmark confirms the connection

API Key Configuration

Some organizations connect Document360 using an API key rather than OAuth. In this case, your organization administrator provides the API key through the admin panel. Individual users do not need to perform any setup.

Disconnecting

To disconnect:

1. Go to Settings > Integrations
2. Click Disconnect next to Document360
3. Confirm the disconnection

Existing context blocks referencing the disconnected account will display a connection error.

Context Blocks

Document360 context blocks inject knowledge base content into your thread context. The AI receives the full article text and can reference it in every response.

Adding Document360 Context

1. Open Manage Context in your thread
2. Go to the Add New tab
3. Click Document360
4. Browse the category tree or search for articles by title or keyword
5. Select articles or categories to include
6. Click Add Context

Articles vs. Categories

Document360 content is organized into categories (folders) containing articles. When adding context, you can select either:

Articles -- Individual knowledge base articles. Each article appears with a blue badge in the context list. The AI receives the full article content including any structured sections, tables, and code blocks.

Categories -- Entire documentation sections. Each category appears with an amber badge in the context list. When a category is selected, all articles within that category (and its subcategories) are included. This is useful when the AI needs comprehensive coverage of a topic area.

Selection Type	Badge Color	Content Included
Article	Blue	Single article content
Category	Amber	All articles in category and subcategories

Cache Duration

Document360 content is cached after the initial fetch. Configure the cache duration based on how frequently your documentation changes:

Duration	Use Case
1 hour	Actively maintained documentation with frequent updates
24 hours	Stable documentation (default)
7 days	Reference documentation that rarely changes

To adjust, edit the context block and modify the Cache Duration setting.

Content Formatting

Article content is delivered to the AI in clean Markdown format. Document360's rich content elements are preserved:

• Headings and section structure
• Ordered and unordered lists
• Tables
• Code blocks with language syntax
• Callouts and admonitions
• Internal links (resolved to article titles)

Images and embedded media are excluded from the text representation, as they cannot be processed as text context.

Available Tools

Document360 tools allow the AI to search and retrieve knowledge base content dynamically during a conversation.

Enabling Document360 Tools

1. Open your thread
2. Click the Tools button in the toolbar
3. Search for "Document360" in the Available Tools grid
4. Toggle the Document360 tools on

Tool Capabilities

Tool	Description
Article Search	Search articles by keyword, title, or content across the entire knowledge base
Category Browse	List categories and their articles for structured navigation
Article Read	Fetch the full content of a specific article by ID or slug

How Tools Differ from Context Blocks

• Context blocks provide the AI with persistent access to specific articles throughout the conversation. The content is always present in the prompt.
• Tools let the AI search for and retrieve articles on demand. This is more token-efficient when you do not know in advance which articles are relevant.

For broad topic coverage, use a category context block. For targeted, on-demand lookups, use tools.

Use Cases

Support Automation

Ground AI responses in your official documentation. Attach relevant product documentation categories as context and instruct the AI to answer customer questions using only the provided knowledge base content.

Example prompt: "Using the attached product documentation, draft a response to this customer question: How do I configure SSO with SAML?"

Knowledge Retrieval

Enable Document360 tools to let the AI search your knowledge base during a conversation. This is useful when handling diverse questions that span multiple documentation areas.

Example prompt: "Search our knowledge base for articles about API rate limits and summarize the current policies."

Documentation-Grounded Responses

Combine Document360 context with other data sources. Attach product documentation alongside a customer's support ticket to generate responses that reference specific articles and provide accurate technical guidance.

Example prompt: "Review this support ticket and draft a response. Reference specific articles from our knowledge base where applicable."

Onboarding Content

Load onboarding documentation into a thread to create an interactive training assistant. New team members can ask questions and receive answers grounded in your official procedures and guidelines.

Documentation Gap Analysis

Attach a category of documentation and ask the AI to identify missing topics, inconsistencies, or outdated content based on a list of expected coverage areas.

Example prompt: "Review the API documentation category. Compare against this list of endpoints and identify any that are not documented."

Best Practices

1. Use categories for breadth, articles for depth -- When the conversation covers a broad topic, attach the category. When it focuses on a specific procedure, attach individual articles.
2. Combine with system prompts -- Set a system prompt instructing the AI to cite documentation sources and only use information from the provided knowledge base.
3. Set appropriate cache durations -- Match the cache to your documentation update frequency. Actively maintained docs benefit from shorter caches.
4. Monitor token usage -- Categories with many articles can consume significant tokens. Check the context usage indicator when attaching large categories.
5. Use tools for unpredictable queries -- When you cannot anticipate which articles will be relevant, enable Document360 tools instead of pre-loading context.

Troubleshooting

Issue	Solution
"Connection expired" error	Re-authorize Document360 in Settings > Integrations
Articles not appearing in search	Verify the articles are published (draft articles are not indexed)
Empty article content	Check that the article has content in the connected project version
Category showing no articles	Verify the category contains published articles in Document360

Related Documentation

• Context Blocks -- Full context block documentation
• Tools Overview -- How tools work in GPT Workbench
• MCP Tools -- Model Context Protocol tools
• Teams -- Team management and collaboration