Do Copilot Agents have trouble reading through tables and charts in Sharepoint articles?

Anonymous
2025-07-22T22:17:05.3166667+00:00

If I'm creating a Copilot agent with knowledge sources that consist of many articles, will the agent be able to differentiate between similar terms? I'm having issues with the agent responding incorrectly if the client asks for information similar to the following example. If the client asks the agent:

What are the steps to registering for GitHubCopilot?
It connects to articles that talk about GitHubCopilot and Copilot.

Also, is the Copilot agent able to read through tables and comparison charts?

Microsoft Copilot | Microsoft 365 Copilot | Development

Answer accepted by question author
Karan Shewale 2,585 Reputation points Microsoft External Staff
2025-07-23T10:23:18.17+00:00

Hi Shahana,

Great questions about Copilot Studio's knowledge source capabilities! Let me address both your concerns:

Table and Chart Reading Capabilities

  • Current Limitations:
    • Tables: Copilot agents can read text-based tables but struggle with complex formatting
    • Charts/ImagesCannot read visual charts, graphs, or image-based content
    • Complex layouts: May lose context in multi-column or nested table structures
  • What Works:
    • Simple HTML/markdown tables
    • CSV-style data in text format
    • Structured lists and bullet points
  • What Doesn't Work Well:
    • Excel embedded tables
    • Image-based charts/graphs
    • Complex SharePoint list views
    • Visually formatted comparison charts
    Term Differentiation Issues Your "GitHubCopilot" vs "Copilot" example highlights a common problem:
  • Why This Happens:
    • Semantic similarity: AI treats "GitHubCopilot" and "Copilot" as related terms
    • Partial matching: Knowledge search finds articles containing both terms
    • Context bleeding: Similar terminology across different articles creates confusion
    Solutions and Workarounds
  • 1. Improve Knowledge Source Structure For better term differentiation: Article Titles:  "Copilot Setup Guide"  "GitHub Copilot Setup Guide - Developer Tool"  "Using Copilot"    "Using Microsoft Copilot Studio - Agent Builder"
  • Add clear context in content:
    • Use full product names consistently
    • Add disambiguation in first paragraph
    • Include product categories/contexts
  • 2. Enhance Table Content for AI Reading Instead of complex tables, use: ## GitHub Copilot Registration Steps:
  • 1. Step 1: Account Setup    - Navigate to GitHub.com    - Sign in to your account    
  • 2. Step 2: Subscription    - Go to GitHub Copilot settings    - Choose individual or business plan Rather than:

StepActionDetails1SetupNavigate...- 3. Use Structured Content Format

For comparison data:

## Product Comparison: GitHub Copilot vs Microsoft Copilot Studio

### GitHub Copilot (Developer Tool):

- Purpose: Code completion and generation

- Platform: IDE integration

- Cost: $10/month individual

### Microsoft Copilot Studio (Agent Builder):

- Purpose: Chatbot and agent creation

- Platform: Web-based builder

- Cost: Per conversation pricing

  • 4. Implement Content Tagging Add metadata to your SharePoint articles:
    • Tags: "github-copilot", "microsoft-copilot-studio"
    • Categories: "developer-tools", "agent-building"
    • Product-specific keywords in article metadata
  • 5. Knowledge Source Organization Separate knowledge sources by product:
    • Source 1: "GitHub Copilot Documentation"
    • Source 2: "Microsoft Copilot Studio Guides"
    • Source 3: "General AI Copilot Information"
    This helps the agent understand context boundaries. Alternative Approaches
  • For Complex Tables/Charts:
    1. Convert to structured text before adding to knowledge base
    2. Use Power BI integration for data visualization queries
    3. Create text summaries of chart information
    4. Add table descriptions explaining what the data shows
  • For Term Disambiguation:
    1. Create a glossary article defining different Copilot products
    2. Use prompt engineering in your agent instructions:"When users ask about 'Copilot', clarify which product they mean:- GitHub Copilot (code assistant)- Microsoft Copilot Studio (agent builder)- Microsoft 365 Copilot (productivity assistant)"
  • Best Practices Content Structure:
    • Use clear headings with product names
    • Repeat key terms throughout content
    • Add context clues in every section
  • Knowledge Base Design:
    • Separate similar products into different sources
    • Use consistent naming conventions
    • Include disambiguation content
  • Agent Instructions:
    • Train the agent to ask clarifying questions
    • Set up confirmation flows for ambiguous requests
    The key is making your content AI-friendly through clear structure and explicit context, rather than relying on visual formatting that the agent can't interpret. Would you like specific guidance on restructuring your SharePoint content or setting up disambiguation flows in your agent?

Thanks,  

Karan Shewale. 

*************************************************************************  

If the response is helpful, please click "Accept Answer" and upvote it. You can share your feedback via Microsoft Teams Developer Feedback link. Click here to escalate. 

Was this answer helpful?

0 comments No comments

0 additional answers

Sort by: Most helpful

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.