7 min read

EvidAI Chat

Your 24/7 AI research methodology expert and assistant

EvidAI Chat

EvidAI Chat is your intelligent assistant for systematic review methodology—available anywhere in the platform to answer questions, provide guidance, and help you navigate complex research decisions.


Overview

Think of EvidAI Chat as having an experienced systematic review methodologist available 24/7. It understands research methods, knows the platform, and can help with everything from basic questions to complex methodological decisions.

CapabilityExamples
Methodology HelpStatistical questions, tool selection
Platform GuidanceHow to use EvidAI features
Workflow SupportNext steps, best practices
TroubleshootingResolving issues and errors

What EvidAI Chat Can Help With

Methodology Questions

TopicExample Questions
Statistics"What model should I use with high heterogeneity?"
Quality Assessment"When do I use RoB 2 vs ROBINS-I?"
Synthesis"How do I handle studies with different outcome measures?"
Reporting"What should I include in my PRISMA flow diagram?"

Platform Navigation

  • "How do I start a new systematic review?"
  • "Where can I find my exported data?"
  • "How do I invite team members?"
  • "What's the difference between Compose and Evidence Studio?"

Workflow Guidance

  • "What should I do after screening is complete?"
  • "How do I handle disagreements between reviewers?"
  • "What's the best way to organize my extracted data?"
  • "How long should protocol development take?"

Using EvidAI Chat

Accessing Chat

Multiple Access Points:

LocationHow to Access
Floating ButtonClick chat icon in bottom-right
HeaderClick "EvidAI Chat" in navigation
KeyboardPress Ctrl/Cmd + /
Context MenuRight-click → "Ask EvidAI"

Conversation Features

Chat Capabilities:

  • Natural language understanding
  • Context awareness (knows your current task)
  • Conversation memory (remembers earlier discussion)
  • Multi-turn dialogues (follow-up questions)

Example Conversation

You: My meta-analysis shows I² of 85%. What should I do?

EvidAI Chat: An I² of 85% indicates substantial heterogeneity. 
Here are recommended steps:

1. **Explore Sources**: 
   • Check for clinical heterogeneity (different populations, interventions)
   • Review methodological heterogeneity (study designs, quality)
   
2. **Conduct Subgroup Analyses**:
   • By study design
   • By population characteristics
   • By intervention variations
   
3. **Consider Your Model**:
   • Random effects is appropriate with this heterogeneity
   • Report prediction intervals alongside confidence intervals
   
4. **Sensitivity Analyses**:
   • Exclude outlier studies
   • Analyze by quality level

Would you like help setting up subgroup analyses in EvidAI?

Context-Aware Assistance

Understanding Your Context

EvidAI Chat knows:

ContextHow It Helps
Current PageOffers relevant suggestions
Active ReviewReferences your specific data
Workflow StageAppropriate guidance for phase
Recent ActionsUnderstands what you just did

Proactive Suggestions

The chat may offer help when it detects:

  • Potential issues with your approach
  • Opportunities to use features
  • Common next steps
  • Best practices to apply

Best Practices

Getting Good Answers

  1. Be Specific: Detailed questions get better answers
  2. Provide Context: Mention your specific situation
  3. Ask Follow-Ups: Dig deeper when needed
  4. Use Proper Terms: Research terminology helps

When to Use Chat

Good ForBetter Alternative
Quick methodology questionsDeep literature review (use EvidAsk)
Platform how-to questionsFormal documentation (use Docs)
Workflow guidanceComplex analysis (consult statistician)
TroubleshootingTechnical bugs (contact support)

Remember: EvidAI Chat is a helpful assistant, but important methodological decisions should be discussed with your team and domain experts.


Integration

EvidAI Chat connects across the platform:

  • Direct Actions: "Show me my included studies" → navigates to list
  • Feature Explanations: "What does this button do?" → contextual help
  • Workflow Triggers: "Start quality assessment" → begins process
  • Documentation Links: References official documentation
Did this article help?
Still stuck?