AI Intelligent Search
Find information faster with AI-powered search refinements and natural language queries.
What is Intelligent Search?
Intelligent Search uses AI to understand what you're looking for and help you find it faster:
- Natural Language Queries: Search like you talk ("contracts signed in 2024")
- Query Interpretation: AI understands your intent
- Smart Refinements: Suggestions to narrow or expand your search
- Context-Aware: Understands document types, dates, relationships
- Learning: Improves based on your search patterns
Why Use It?
Find Information Faster:
- No need to remember exact keywords
- AI suggests better search terms
- Understands synonyms and related concepts
Better Results:
- More relevant documents ranked higher
- Filters suggested automatically
- Related documents surfaced
Save Time:
- Fewer searches needed to find what you want
- No trial-and-error with keywords
- Instant query refinements
How It Works
Basic Search Flow
1. You enter query: "contracts signed in 2024"
↓
2. AI interprets: Looking for legal contracts signed during 2024
↓
3. AI suggests refinements:
- Contracts signed in Q1 2024
- Contracts with specific parties in 2024
- Unsigned contracts from 2024
↓
4. You select refinement or use original query
↓
5. Results displayed with AI-enhanced ranking
Query Interpretation
AI analyzes your query to understand:
- Intent: What are you trying to find?
- Entity Types: Documents, people, dates, amounts?
- Time Range: Specific dates or periods?
- Document Types: Contracts, invoices, emails?
- Relationships: Connections between entities?
Example Interpretations:
| Your Query | AI Interpretation |
|---|---|
| "employment contract jane" | Employment contracts mentioning Jane |
| "invoices > $10000 last month" | Invoices over $10,000 from previous month |
| "emails urgent yesterday" | Urgent emails received yesterday |
| "payments to acme 2024" | Payment documents to Acme Corp in 2024 |
Using Intelligent Search
Natural Language Queries
Instead of This (keyword search):
contract AND (2024 OR 2023) AND (TechCorp OR "Tech Corp")
Try This (natural language):
TechCorp contracts from 2024
More Examples:
| Natural Language | What It Finds |
|---|---|
| "show me all invoices from last quarter" | Invoices from Q3/Q4 2024 |
| "contracts expiring soon" | Contracts expiring in next 90 days |
| "documents about layoffs" | Documents mentioning layoffs/terminations |
| "emails from john marked urgent" | Urgent emails from John |
| "financial reports Q3" | Financial reports from Q3 |
Query Refinements
AI suggests ways to improve your search:
Too Many Results?
Query: "contracts"
→ 543 results
AI Suggestions:
- Contracts signed in 2024 (127 results)
- Employment contracts (89 results)
- Contracts expiring in 2025 (34 results)
Too Few Results?
Query: "employment agreement jane smith techcorp 2024"
→ 0 results
AI Suggestions:
- Employment agreements in 2024 (12 results)
- Contracts mentioning Jane Smith (3 results)
- TechCorp documents (45 results)
Ambiguous Query?
Query: "contracts"
AI Interpretation:
"You might be looking for:"
- Active contracts (234 results)
- Expired contracts (187 results)
- Draft contracts (45 results)
- All contracts (543 results)
Document Type Filtering
AI recognizes document types in your query:
Examples:
- "invoices from Q1" → Automatically filters by document type: invoice
- "legal filings" → Filters by category: legal
- "email about payment" → Filters by type: correspondence
Supported Types:
- Contracts
- Invoices
- Legal documents
- Financial reports
- Correspondence (emails, letters)
- Personal documents
Date and Time Queries
AI understands relative and absolute dates:
Relative Dates:
- "yesterday", "last week", "last month", "last quarter", "last year"
- "this week", "this month", "this quarter", "this year"
- "next week", "next month" (for future dates like expiration)
Absolute Dates:
- "January 2024", "Q1 2024", "2024"
- "January 1, 2024", "01/01/2024"
- "between January 1 and March 31"
Date Ranges:
- "2023-2024" → Documents from 2023 or 2024
- "last 6 months" → Documents from past 6 months
- "expiring in 30 days" → Documents expiring in next 30 days
Amount and Number Queries
Search by monetary amounts or numbers:
Examples:
- "invoices over $10,000" → Invoices with amount > $10,000
- "contracts worth $50k-$100k" → Contracts between $50K-$100K
- "payments under $1000" → Small payments
Supported Comparisons:
- Greater than: "over $X", "> $X", "more than $X"
- Less than: "under $X", "< $X", "less than $X"
- Equal to: "$X", "exactly $X"
- Ranges: "$X to $Y", "$X-$Y"
Advanced Search Features
Entity-Based Search
Search by extracted entities:
By Name:
"documents mentioning jane smith"
→ All documents with "Jane Smith" entity
By Organization:
"contracts with techcorp"
→ All contracts where TechCorp is a party
By Location:
"documents related to california"
→ Documents mentioning California
Relationship Search
Find documents based on connections:
Examples:
- "documents linking jane and john" → Docs mentioning both
- "contracts between techcorp and acme" → Contracts with both parties
- "payments from techcorp to john" → Payment flow between entities
Semantic Search
AI understands concepts, not just keywords:
Query: "employee termination" Also Finds:
- "layoff", "dismissed", "fired", "let go"
- "separation agreement", "exit interview"
- "final paycheck", "severance"
Query: "data breach" Also Finds:
- "security incident", "unauthorized access"
- "hacked", "compromised", "leaked"
- "cyber attack", "intrusion"
Search Results
AI-Enhanced Ranking
Results ranked by relevance:
Ranking Factors:
- Query Match: How well document matches your query
- Recency: Newer documents ranked higher (configurable)
- Document Type: Matches your implied type (contracts, invoices, etc.)
- Entity Matches: Documents with matching entities
- User History: Documents you've accessed before
- Confidence: AI's confidence in the match
Result Preview
Each result shows:
- Document Title: Name of document
- Summary: AI-generated summary (if available)
- Key Points: Relevant excerpts matching query
- Metadata: Type, date, size, entities
- Confidence Score: How confident AI is this matches
Example Result:
📄 Employment Agreement - Jane Smith.pdf
Summary: Employment contract for Senior Software Engineer position
at TechCorp, effective January 15, 2024. Annual salary $150,000.
Key Matches:
- "Jane Smith" (Employee)
- "TechCorp" (Employer)
- Date: 2024-01-15 (Start date)
- Amount: $150,000 (Salary)
Confidence: 0.95 (High)
Type: Contract | Size: 245 KB | Uploaded: Jan 10, 2024
Export Results
Export search results:
- CSV: Spreadsheet with metadata
- PDF: Report with summaries
- JSON: Programmatic access
- Bulk Download: Download all matching documents
Practical Examples
Example 1: Finding Expired Contracts
Goal: Find contracts that expired in 2024 and need renewal.
Query: "contracts expired in 2024"
AI Interpretation: "Looking for contracts with expiration dates in 2024"
Results: 23 contracts Suggested Actions:
- Renew high-value contracts first
- Contact parties to negotiate renewals
- Archive contracts not renewing
Example 2: Financial Audit
Goal: Find all invoices over $10,000 from Q1 2024 for audit.
Query: "invoices over $10,000 Q1 2024"
AI Interpretation: "Invoices with amounts exceeding $10,000 from January-March 2024"
Results: 15 invoices Export: CSV with invoice numbers, amounts, dates, vendors
Example 3: Legal Discovery
Goal: Find all documents mentioning specific parties for legal case.
Query: "documents mentioning Acme Corp and John Smith 2023-2024"
AI Interpretation: "All documents from 2023-2024 that mention both Acme Corp and John Smith"
Results: 34 documents (contracts, emails, invoices, correspondence) Suggested Refinements:
- Contracts only (12 documents)
- Correspondence only (18 documents)
- Financial documents only (4 documents)
Best Practices
Writing Effective Queries
✅ Good Queries:
- "employment contracts expiring in 2025"
- "invoices from acme corp last quarter"
- "urgent emails about security incident"
- "financial reports Q3 2024"
❌ Avoid:
- Too vague: "documents" (too broad)
- Boolean operators: "contract AND (2024 OR 2023)" (unnecessary with AI)
- Exact matches only: Quotes force exact match, may miss relevant docs
Tips:
- Be specific about what you're looking for
- Include timeframes when relevant
- Mention document types if known
- Use natural language, not boolean logic
Using Filters
Combine AI search with manual filters:
Filters Available:
- Document Type: Contract, Invoice, Legal, etc.
- Date Range: Custom date ranges
- File Type: PDF, Word, Image, etc.
- Vault: Search specific vaults only
- Tags: Documents with specific tags
- Size: File size ranges
Example:
Query: "contracts"
+ Filter: Date range = Last 12 months
+ Filter: Vault = "Corporate Contracts"
+ Filter: Tag = "Active"
→ Active corporate contracts from last year
Refining Searches
Start Broad, Then Narrow:
- Start with general query: "contracts"
- Review AI refinement suggestions
- Select refinement: "Contracts expiring in 2025"
- Apply additional filters if needed
Iterate Based on Results:
- Too many results? Add more specific terms or filters
- Too few results? Remove filters or broaden query
- Wrong results? Try suggested refinements
Troubleshooting
No Results Found
Solutions:
- Broaden query: Remove specific terms
- Check spelling: AI helps but may miss misspellings
- Remove filters: Date/type filters may be too restrictive
- Try synonyms: Different terminology for same concept
- Search all vaults: May be in different vault
Irrelevant Results
Solutions:
- Add specificity: "contracts TechCorp 2024" instead of "contracts"
- Use refinements: AI-suggested refinements often improve relevance
- Apply filters: Document type, date range, vault filters
- Quote exact phrases: "employment agreement" matches exact phrase
- Exclude terms: "-draft" excludes documents with "draft"
Slow Search
Solutions:
- Narrow scope: Search specific vault instead of all vaults
- Use filters: Reduce search space with date/type filters
- Check indexing: New documents may still be indexing
- Clear cache: Settings → Clear Search Cache
FAQ
Does Intelligent Search work without AI?
Yes. If AI is disabled, search reverts to traditional keyword search. AI features (interpretation, refinements) are unavailable.
Can I disable AI search suggestions?
Yes. Settings → AI Preferences → Disable "Search Refinements"
You'll still get search results, but without AI-powered refinements.
Does search work across all vaults?
Yes, by default. You can filter to search specific vaults only.
How is search data used?
Search queries are used to:
- Improve search results for you
- Train AI models (anonymized, aggregated data only)
- Provide better refinement suggestions
Opt-out: Settings → Privacy → Disable "Search Analytics"
Can I see my search history?
Yes. Settings → Search History shows recent searches.
- Export search history (CSV, JSON)
- Clear search history anytime
- Disable search history tracking
Does search work with encrypted documents?
Yes. Documents are decrypted for search indexing, then re-encrypted. Search index is also encrypted.
Can I search by document content?
Yes. Full-text search indexes all document content (text, OCR from images).
How often is the search index updated?
Real-time. New documents indexed within 1-5 minutes of upload.
API Access
Search Documents via API
curl -X POST https://api.torvussecurity.com/v1/ai/search \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"query": "contracts expiring in 2025",
"vaultId": "vault_xyz789",
"maxResults": 10
}'
Response:
{
"query": "contracts expiring in 2025",
"interpretation": "Looking for contracts with expiration dates in 2025",
"suggestedRefinements": [
"Contracts expiring in Q1 2025",
"High-value contracts expiring in 2025",
"Contracts expiring in next 90 days"
],
"results": [
{
"documentId": "doc_abc123",
"title": "Service Agreement - Acme Corp.pdf",
"summary": "Service agreement expiring March 31, 2025",
"confidence": 0.95,
"matchedEntities": [
{"type": "date", "value": "2025-03-31", "context": "Expiration date"}
]
}
],
"totalResults": 23,
"confidence": 0.92
}
for full documentation.
Getting Help
Support:
- Search Issues: search-support@torvussecurity.com
- AI Problems: ai-support@torvussecurity.com
- Feature Requests: feature-requests@torvussecurity.com
Resources:
Last Updated: October 15, 2025