Intelligent Import (Pattern-Based)
Import requirements and FMEA data from Word and Excel documents using intelligent pattern recognition - no AI required, your data stays completely private.
Intelligent Import uses pattern matching and heuristics to extract requirements from documents. Your data never leaves your browser or server - no AI APIs are called.
AI Import (Advanced tier) uses Google Gemini to interpret natural language. More powerful for unstructured documents, but sends data to external AI service.
Overview
The Intelligent Import Wizard analyzes your documents using:
- Pattern Recognition - Detects numbered lists, headings, table structures
- Column Synonym Matching - Recognizes 50+ common column name variations
- INCOSE Best Practices - Validates requirements against industry standards
- Confidence Scoring - Shows you how reliable the extraction is
Availability
| Tier | Word Import | Excel/FMEA Import |
|---|---|---|
| Free | ❌ | ❌ |
| Basic ($49/mo) | ✅ | ✅ |
| Advanced ($99/mo) | ✅ | ✅ |
| Power ($299/mo) | ✅ | ✅ |
| Enterprise | ✅ | ✅ |
Intelligent Import is available starting at Basic tier because:
- No AI compute costs (no external API calls)
- Data stays private (important for regulated industries)
- Provides excellent value for document conversion
Supported Formats
Word Documents (.docx)
The wizard detects requirements using multiple strategies:
| Detection Strategy | Confidence | Best For |
|---|---|---|
| Tables | 95% | Structured requirement tables |
| Headings | 85% | Section-based documents |
| Numbered Lists | 80% | Hierarchical numbered requirements |
| "Shall" Statements | 70% | Narrative text with requirements |
Works best with:
- Tables with ID, Description, Type columns
- Numbered sections (1.0, 1.1, 1.1.1)
- Heading-based organization (H1, H2, H3)
- "Shall" statement patterns
Excel Files (.xlsx)
The wizard intelligently maps columns using synonym matching:
Requirement Columns Recognized:
| Standard Name | Synonyms Automatically Detected |
|---|---|
| ID | custom_id, req_id, requirement_id, number, # |
| Text | description, statement, requirement, body, content |
| Type | category, classification, req_type, level |
| Status | state, phase, lifecycle |
| Parent ID | parent, parent_id, parent_custom_id, hierarchy |
FMEA Columns Recognized:
| Standard Name | Synonyms Automatically Detected |
|---|---|
| Function | item_function, purpose |
| Failure Mode | potential_failure, failure, mode |
| Effect | failure_effect, consequence, impact |
| Cause | failure_cause, root_cause, reason |
| Severity | sev, s |
| Occurrence | occ, o, frequency |
| Detection | det, d, detectability |
How to Use
Step 1: Access the Wizard
- Open your project
- Click Project → Import with Wizard
- The Import Wizard modal opens
Step 2: Select Import Type
Choose what you want to import:
- Requirements - From Word or Excel
- FMEA Data - From Excel with S/O/D ratings
Step 3: Upload Your File
- Click Choose File or drag-and-drop
- Supported:
.docx,.xlsx - Maximum file size: 10 MB
Step 4: Review Analysis
The wizard analyzes your file and shows:
For Word Documents:
Detection Strategy: Tables (95% confidence)
├── Found: 3 tables with requirement-like structure
├── Columns: ID, Description, Type, Rationale
└── Estimated requirements: 47
For Excel Files:
Sheet: "Requirements" (Detected: requirements)
├── Mapped Columns:
│ ├── custom_id ← "Req ID" (synonym match)
│ ├── text ← "Description" (synonym match)
│ └── type ← "Category" (synonym match)
└── Rows to import: 156
Step 5: Adjust Mappings (if needed)
If columns aren't mapped correctly:
- Click the dropdown for unmapped columns
- Select the correct target field
- Preview updates automatically
Step 6: Review INCOSE Compliance
The wizard checks your requirements against INCOSE best practices:
Quality Indicators:
- ✅ Uses "shall" for requirements
- ✅ Specific and measurable
- ⚠️ Contains vague terms ("good", "fast")
- ❌ Missing subject or object
Common Issues Flagged:
| Issue | INCOSE Rule | Example |
|---|---|---|
| Missing "shall" | R1 | "The system provides..." → "The system shall provide..." |
| Vague terms | R7 | "fast response" → "response within 100ms" |
| Escape clauses | R8 | "as applicable" → Specify the condition |
| Multiple requirements | R9 | Split into separate statements |
Step 7: Execute Import
- Review the summary showing:
- Total items to import
- Confidence score
- Any warnings
- Click Import
- Requirements appear in your project
Confidence Scoring
The wizard provides transparency about extraction quality:
| Score | Meaning | Recommendation |
|---|---|---|
| 90-100% | Excellent | Import with confidence |
| 80-89% | Good | Review flagged items |
| 70-79% | Fair | Manual review recommended |
| Below 70% | Low | Consider AI Import for better results |
If confidence is below 70%, the wizard suggests trying AI Import (Advanced tier) which can better interpret unstructured or ambiguous documents.
Word Document Best Practices
Recommended Document Structure
Best: Tables with clear columns
| ID | Requirement | Type | Rationale |
|----|-------------|------|-----------|
| SYS-001 | The system shall... | Functional | Safety |
Good: Numbered sections
1.0 System Requirements
1.1 The system shall provide real-time monitoring
1.2 The system shall alert users within 5 seconds
Acceptable: Heading-based
## 2.0 Performance Requirements
### 2.1 Response Time
The system shall respond within 100ms.
What to Avoid
❌ Unstructured narrative - No clear requirement boundaries ❌ Mixed formatting - Some tables, some bullets, some paragraphs ❌ Embedded images - Text in images cannot be extracted ❌ Complex nested tables - May not parse correctly
Excel Best Practices
Recommended Format
| Req ID | Description | Category | Status | Parent |
|---|---|---|---|---|
| SYS-001 | System shall monitor... | System | Draft | |
| SYS-001.1 | Subsystem shall... | Subsystem | Draft | SYS-001 |
Tips:
- Use consistent column names
- Include Parent ID for hierarchy
- First row must be headers
- No merged cells
- No formulas in data columns
FMEA Excel Format
| Function | Failure Mode | Effect | Cause | S | O | D |
|---|---|---|---|---|---|---|
| Monitor temperature | Sensor drift | Incorrect reading | Calibration error | 7 | 4 | 5 |
Required columns for FMEA:
- Function or Item
- Failure Mode
- Effect
- Severity (1-10)
- Occurrence (1-10)
- Detection (1-10)
Comparison with Other Import Methods
| Feature | Intelligent Import | Manual Import | AI Import |
|---|---|---|---|
| Input Format | Word, Excel | Excel, JSON | Word |
| Tier Required | Basic ($49/mo) | Free | Advanced ($99/mo) |
| Data Privacy | ✅ 100% Private | ✅ 100% Private | ⚠️ Sent to AI API |
| Best For | Semi-structured docs | Pre-formatted data | Unstructured narrative |
| Accuracy | 85-95% | 100% | 90-95% |
| Processing | Instant | Instant | 1-3 minutes |
| INCOSE Validation | ✅ Yes | ❌ No | ❌ No |
| Column Detection | ✅ Automatic | ❌ Manual | N/A |
Troubleshooting
"No requirements detected"
Causes:
- Document has no recognizable patterns
- Text is in images or scanned
- Unusual formatting
Solutions:
- Restructure document with clear sections
- Use tables with headers
- Try AI Import for complex documents
"Low confidence score"
Causes:
- Mixed document structure
- Ambiguous column names
- Non-standard formats
Solutions:
- Manually adjust column mappings
- Split document into cleaner sections
- Consider AI Import for interpretation
"Column not mapped"
Causes:
- Unusual column name
- Column name in different language
- Abbreviations not recognized
Solutions:
- Use dropdown to manually map
- Rename columns in source file
- Add common synonyms in future update
"FMEA import missing ratings"
Causes:
- S/O/D columns not detected
- Ratings not in 1-10 scale
- Text values instead of numbers
Solutions:
- Ensure columns are named "Severity", "Occurrence", "Detection"
- Convert text ratings to numbers (High=9, Medium=5, Low=2)
- Use S, O, D abbreviations
Frequently Asked Questions
Is my data sent to any external service?
No. Intelligent Import runs entirely within NirmIQ. Your document is processed on our servers and never sent to third-party AI services like OpenAI or Google.
What's the difference between this and Manual Import?
Manual Import requires you to pre-format an Excel file with exact column names. Intelligent Import analyzes your existing documents and maps columns automatically using synonym matching.
What's the difference between this and AI Import?
Intelligent Import uses pattern matching (rules-based). AI Import uses Google Gemini to interpret natural language. AI Import is better for unstructured documents but requires Advanced tier and sends data externally.
Can I import FMEA and Requirements from the same file?
Currently, you need to import them separately. If your Excel has both on different sheets, run the wizard twice selecting different sheets.
What INCOSE rules are checked?
The wizard validates against key INCOSE Guide for Writing Requirements rules:
- R1: Use "shall" for mandatory requirements
- R7: Avoid vague terms (good, fast, user-friendly)
- R8: Avoid escape clauses (as applicable, if necessary)
- R9: One requirement per statement
- R17: Complete sentences with subject and predicate
Can I customize the synonym dictionary?
Not currently. If you have industry-specific terms that should be recognized, contact support and we can add them.
Related Documentation
- Import Methods Comparison - Compare all import options
- AI-Powered Import - For unstructured documents
- Manual Import Guide - Pre-formatted Excel/JSON
- ReqIF Import - Enterprise tool integration
- INCOSE Best Practices - Requirements writing standards
Need Help?