The MCP Server enables external AI assistants to interact with your Document360 knowledge base in a structured way. These interactions do not bypass your governance model - they operate within your existing version, language, and visibility configuration. Below are practical scenarios that show how different roles can use MCP in their daily documentation workflows.
Technical writers
Problem: Researching, drafting, and refining documentation takes context-switching and manual effort
Technical writers often move between multiple articles, versions, and categories to gather context before writing or updating content. They also need to ensure consistency in tone, terminology, and structure, especially when a Style Guide is configured. Switching between the editor, search, previous versions, and reference articles slows down drafting and review.
How MCP helps
With an AI assistant connected to Document360 through MCP, writers can:
- Ask the assistant to fetch relevant articles before drafting new content
- Brainstorm ideas grounded in existing documentation
- Refine or rewrite content directly within the documentation system
- Update articles without manually copying content between tools
- Move updated articles to the appropriate workflow status and assign reviewers without switching to the portal
- Publish approved content directly from the AI assistant once the review is complete
Instead of drafting in isolation and pasting content later, writers can work with context pulled from their own knowledge base. When updates are made, existing workflow and versioning behavior remains unchanged. Any forked revisions or review states continue to follow normal documentation processes.
Outcome
Technical writers reduce repetitive research, maintain better consistency across articles, and draft and publish content faster, while staying within the same documentation governance model.
Developers
Problem: Developers need version-specific documentation while building features
Developers frequently switch between IDEs and AI-assisted coding environments such as Cursor or other AI-enabled tools. When referencing documentation, they need accurate, version-aware information without manually navigating the knowledge base.
How MCP helps
By connecting Document360 through MCP to their AI assistant within the IDE:
- Developers can retrieve full article content instead of partial snippets
- Search across all project versions by default, or specify a version
- Access documentation in a selected language
- Generate draft documentation for new endpoints or integrations
This allows documentation lookup and content creation to happen directly within the development workflow.
Outcome
Developers spend less time switching contexts and more time building with documentation that reflects the correct version and structure.
Product managers and consultants
Problem: Validating documentation across releases is manual and fragmented
Product managers often need to compare feature descriptions across versions, prepare release notes, or validate that documentation reflects the latest product changes. Manually checking multiple versions and categories can be slow and inconsistent.
How MCP helps
Through MCP-enabled AI assistants, product teams can:
- Retrieve documentation by specific project version
- Compare how features are described across releases
- Identify gaps before launch
- Generate structured summaries for internal alignment or onboarding
Because the interaction is structured, content is retrieved in context rather than as disconnected web output.
Outcome
Documentation becomes easier to validate before release, reducing misalignment between product updates and published knowledge.
Support engineers
Problem: Support tickets expose knowledge gaps
When a support ticket arrives, engineers need to quickly determine whether relevant documentation already exists. Searching manually can be time-consuming, especially across multiple versions and languages. In some cases, documentation may not exist at all.
How MCP helps
With MCP connected to an AI assistant:
- Support engineers can paste ticket context and ask the assistant to fetch relevant articles
- Retrieve content by article name or ID
- Search across versions and languages
- Share article URLs directly with customers
If no relevant article is found, then it means a knowledge gap exists in the knowledge base. Engineers can then:
- Draft a new article based on the ticket context
- Update an existing article to improve clarity
- Push structured changes back into Document360
- Move the article to review and assign the appropriate team member before it goes live
- Publish the article directly once it has been reviewed and approved
All write, workflow, and publishing operations continue to respect existing permissions and governance rules.
Outcome
Support teams resolve tickets faster while continuously improving documentation quality — and can now take content from draft to live without leaving their AI assistant.
Documentation managers and content leads
Problem: Coordinating review, approval, and publishing across a large team is time-consuming
Documentation managers overseeing multiple writers need to track which articles are in draft, which are awaiting review, and which are ready to publish. Coordinating this across a team without a centralized view leads to delays, missed reviews, and inconsistent publishing cadences.
How MCP helps
With an AI assistant connected to Document360 through MCP, documentation managers can:
- Check which workflow statuses are configured and how the review pipeline is structured
- Move multiple articles to the appropriate review stage and assign the right team members in a single conversation
- Set due dates on workflow tasks to keep reviews on schedule
- Publish content that has cleared review without switching to the portal
- Unpublish outdated or incorrect content quickly when something needs to be taken offline
Because all workflow and publishing operations follow the same permissions and governance model as the portal, managers retain full control over who can do what.
Outcome
Documentation managers spend less time chasing review status and manually publishing content, and more time ensuring documentation quality and consistency across the team.
Best practices
Follow these best practices when using MCP with AI assistants:
- Use MCP for retrieving source-of-truth documentation instead of manually copying content
- Verify AI-generated content before moving it to review or publishing it to the knowledge base site
- Specify project version or language when precision is required
- Use clear and specific prompts when requesting content or updates
- Ensure that user roles and permissions are configured correctly, especially for workflow and publishing operations
Limitations and considerations
- AI assistants can only access content permitted by your project configuration
- Restricted or unpublished content remains governed by visibility rules
- AI-generated outputs may require validation before use
- The configuration process may vary depending on the AI assistant
- MCP support may vary depending on the AI assistant or developer tool being used (such as IDE-based assistants or experimental clients). Some tools may have limited or evolving support for MCP capabilities.
- If a team account has different portal roles and content roles across categories and articles, MCP applies the lowest-priority role across all categories and articles. For example, if you have Editor access in one category and Reviewer access in another, MCP treats both as Reviewer.
- If a user is assigned two custom roles, one with view permissions for a set of categories and another with update permissions for a different set of categories, the role with view access takes precedence. If you would like the user to update articles via MCP, please ensure that only the role with update permissions is assigned to the user.
- To use workflow tools via MCP, custom portal roles must have the View workflow assignments and Update workflow assignments permissions explicitly enabled. Without these, workflow status changes cannot be performed through MCP.
- For Page categories, there is no standalone Publish permission in custom role settings. If a custom role has View and Update permissions for categories, publishing Page categories via MCP is automatically allowed, consistent with portal behavior.
- Publishing and unpublishing via MCP is not supported for Step-by-step guides.
Troubleshooting
If you encounter issues when using MCP:
- Confirm that MCP is enabled in Document360
- Verify that the MCP Server URL is correct
- Ensure that OAuth authentication is completed successfully
- Check whether your AI assistant supports MCP or custom connectors
If the issue persists, refer to your AI assistant documentation or contact support.