Skip to content

Control 1.8: Information Architecture Review — Portal Walkthrough

Step-by-step portal procedures for reviewing and optimizing the Microsoft 365 information architecture for Copilot readiness.

Prerequisites

  • SharePoint Admin and Global Reader roles
  • Access to SharePoint admin center and Microsoft Purview
  • Content inventory documentation or content management system records
  • Information architecture governance committee or content owners

Steps

Step 1: Inventory Site Architecture

Portal: SharePoint admin center Path: SharePoint admin center > Sites > Active sites

Review the complete site inventory to understand the current information architecture. Categorize sites by: - Type: Team sites, communication sites, hub sites, OneDrive - Purpose: Departmental, project, cross-functional, knowledge base - Hub associations: Which sites are connected to hub sites - Content volume: Storage usage and item counts

Export the site list for documentation and analysis.

Step 2: Evaluate Hub Site Structure

Portal: SharePoint admin center Path: SharePoint admin center > Sites > Active sites > Hub sites filter

Review the hub site hierarchy. Hub sites define content relationships that affect how Copilot discovers and associates content. Verify: - Hub sites align with business unit or function boundaries - Associated sites are correctly linked to their parent hub - No orphaned sites that should be hub-associated

Identify opportunities to improve hub structure for better Copilot content grounding.

Step 3: Review Content Type and Metadata Standards

Portal: SharePoint admin center Path: SharePoint admin center > Content services > Content type gallery

Review the content type hierarchy published from the content type hub. Consistent content types and metadata improve Copilot's ability to understand and classify content accurately.

Verify that FSI-relevant content types are defined (e.g., client agreements, regulatory filings, research reports, policies and procedures).

Step 4: Assess Taxonomy and Term Store

Portal: SharePoint admin center Path: SharePoint admin center > Content services > Term store

Review the managed metadata term store for FSI-relevant term groups: - Business unit taxonomy - Document type classification - Regulatory domain categorization - Geographic region tagging

Verify term sets are current and actively used for content tagging.

Step 5: Map Content to Copilot Use Cases

Document how the current information architecture supports intended Copilot use cases. For each planned Copilot use case, identify: - Which content sources will be accessed - Whether content is properly organized and labeled - Any architectural gaps that could result in poor Copilot responses - Recommended architecture improvements

Step 6: Review Copilot in SharePoint Generated Architecture

Portal: SharePoint site or library experience; Microsoft 365 admin center Paths: SharePoint site or library > Copilot in SharePoint entry points; Microsoft 365 admin center > Copilot > Settings > View all > AI providers operating as Microsoft subprocessors

If Copilot in SharePoint is available in the tenant, review any AI-generated sites, document libraries, public views, autofill metadata columns, and extraction prompts before treating them as authoritative information architecture.

Verify: - Create sites with AI output has an accountable owner, approved sensitivity label, and appropriate hub/site association - AI-created document libraries have reviewed metadata schema, column names, and extraction prompts - Autofill columns are reviewed for accuracy before their metadata is used for Copilot grounding or governance reports - Anthropic model availability and user/group scope match the organization's third-party AI risk posture

FSI Recommendations

Tier Recommendation
Baseline Complete information architecture inventory and document current state
Recommended Optimize hub site structure and implement consistent content types for FSI data
Regulated Formal information architecture governance with annual review; content types and metadata standards enforced via policy; AI-generated SharePoint structures and autofill metadata reviewed before production reliance

Next Steps