Smart Search
Smart Search revolutionizes literature searching by combining AI intelligence with comprehensive database access. Build better searches, find more relevant papers, and never miss important studies.
Overview
Traditional literature searching is tedious and error-prone. Smart Search uses AI to help you build comprehensive search strategies, identify relevant papers, and continuously monitor for new publications.
| Feature | Traditional Search | Smart Search |
|---|---|---|
| Strategy Building | Manual, time-consuming | AI-assisted, guided |
| Database Coverage | One at a time | Multiple simultaneous |
| Result Relevance | Many false positives | AI-ranked by relevance |
| Monitoring | Manual re-runs | Automated alerts |
Search Capabilities
1. AI-Assisted Strategy Building
Build comprehensive searches with AI guidance:
| Step | AI Assistance |
|---|---|
| Concept Identification | Suggests search concepts from your question |
| Term Expansion | Recommends synonyms, MeSH terms, variations |
| Syntax Generation | Creates properly formatted search strings |
| Validation | Checks for common errors and gaps |
Example:
Your Question: "Does exercise improve cognitive function in older adults?"
AI-Generated Concepts:
POPULATION:
• older adults, elderly, aged, seniors, geriatric
• MeSH: "Aged"[Mesh]
INTERVENTION:
• exercise, physical activity, aerobic, resistance training
• MeSH: "Exercise"[Mesh], "Physical Conditioning"[Mesh]
OUTCOME:
• cognition, cognitive function, memory, executive function
• MeSH: "Cognition"[Mesh], "Memory"[Mesh]
2. Multi-Database Search
Search across databases simultaneously:
| Database | Type | Coverage |
|---|---|---|
| PubMed | Biomedical | 35M+ articles |
| Cochrane | Systematic reviews | High-quality SR |
| Embase | Biomedical | European focus |
| CINAHL | Nursing | Allied health |
| PsycINFO | Psychology | Mental health |
| Web of Science | Multidisciplinary | Citation data |
3. AI Relevance Ranking
Results ranked by true relevance:
Beyond Keywords: Smart Search doesn't just match keywords—it understands your research question and ranks results by how well they address your actual needs.
Ranking Factors:
- Topic relevance to your question
- Study design appropriateness
- Population match
- Outcome alignment
- Publication quality signals
Building a Search
Step-by-Step Process
Step 1: Define Your Question
- Enter your research question in natural language
- Or use PICO format for structured input
- AI analyzes and confirms understanding
Step 2: Review Concepts
- AI suggests concept blocks
- Add, remove, or modify concepts
- Expand with synonyms and controlled vocabulary
Step 3: Set Parameters
- Date range filters
- Language restrictions
- Study design filters
- Publication type limits
Step 4: Select Databases
- Choose databases to search
- Review database-specific syntax
- Set deduplication preferences
Step 5: Execute Search
- Run across all selected databases
- View combined, deduplicated results
- Apply AI relevance ranking
Search Strategy Templates
Use pre-built templates for common reviews:
| Template | Optimized For |
|---|---|
| Intervention Review | RCTs, treatment effectiveness |
| Diagnostic Review | Sensitivity/specificity studies |
| Prognostic Review | Cohort studies, risk factors |
| Qualitative Review | Qualitative research methods |
Search Results
Results Interface
| Feature | Description |
|---|---|
| Relevance Scores | AI-assigned relevance rating |
| Quick Preview | See abstract without leaving results |
| Bulk Actions | Select multiple for batch operations |
| Export Options | Multiple format support |
Filtering Results
Narrow results after search:
- By Relevance: High, medium, low confidence
- By Date: Year ranges
- By Study Type: RCT, cohort, etc.
- By Status: Unreviewed, saved, excluded
Saving Results
| Action | Result |
|---|---|
| Save to Library | Add to your document library |
| Add to Review | Include in active systematic review |
| Export | RIS, BibTeX, CSV formats |
| Share | Send results to collaborators |
Search Monitoring
Saved Searches
Set up ongoing monitoring:
Configuration Options:
- Search name and description
- Update frequency (daily, weekly, monthly)
- Notification preferences
- Auto-add to collection
Alert Management
| Alert Type | When |
|---|---|
| New Results | New papers match your search |
| High Relevance | High-confidence matches found |
| Summary Digest | Periodic overview of new results |
Advanced Features
Boolean Builder
Visual interface for complex searches:
- Drag-and-drop concept blocks
- Visual AND/OR/NOT connections
- Real-time syntax preview
- Syntax validation
Citation Searching
Find papers through citations:
- Cited By: Papers citing a known study
- References: Papers cited by a study
- Similar Articles: Content-based similarity
Deduplication
Automatic duplicate handling:
- Cross-database matching
- Title/author/DOI comparison
- Manual review option for uncertain matches
- Deduplication report
Integration
With Screening
Seamlessly move to screening:
- Complete your search
- Review and finalize results
- Click "Start Screening"
- Results imported automatically
With Library Manager
- Save all results to library
- Organize into collections
- Access full-text when available
Export for Documentation
Generate search documentation:
- PRISMA flow diagram data
- Search strategy appendix
- Database coverage report
Best Practices
Building Effective Searches
- Start with Clear Question: Well-defined question = better search
- Be Comprehensive: Include all relevant synonyms and terms
- Test and Iterate: Run preliminary searches, refine strategy
- Document Everything: Record rationale for search decisions
Quality Assurance
| Check | Why |
|---|---|
| Peer Review | Have colleague review strategy |
| Sensitivity Test | Ensure known relevant papers are found |
| Specificity Balance | Not too many irrelevant results |
| Reproducibility | Strategy can be re-run with same results |
Pro Tip: Before finalizing your search, identify 5-10 papers you know should be included and verify your search finds them all.