AI Assistant

AI Assistant

1. Overview

RAILWISE AI Assistant is an intelligent assistant based on large language models (LLM), designed for engineering monitoring scenarios. It supports natural language queries, intelligent data analysis, automatic report generation, and anomaly detection.

2. Features

2.1 Natural Language Query

Query monitoring data using natural language without learning complex query syntax.

Examples:

"Show the settlement trend of point JC-01 in the last 30 days"
"Compare the settlement difference between JC-01 and JC-02"
"Find all points exceeding the warning value"
"Analyze the data change pattern this week"

2.2 Intelligent Data Analysis

AI automatically analyzes monitoring data and identifies anomalies and trends.

Analysis Types:

Analysis Type	Description	Output
Trend Analysis	Analyze data change trends	Trend chart + text description
Anomaly Detection	Identify abnormal data points	Anomaly list + cause analysis
Correlation Analysis	Analyze correlations between multiple points	Correlation matrix + chart
Prediction Analysis	Predict future trends	Prediction chart + confidence interval

2.3 Automatic Report Generation

Automatically generate monitoring reports based on data and templates.

Report Types:

Report Type	Content	Generation Time
Daily Report	Data summary, trend analysis	1-2 minutes
Weekly Report	Weekly summary, evaluation	3-5 minutes
Monthly Report	Monthly summary, stage evaluation	5-10 minutes
Warning Report	Anomaly description, disposal suggestions	1-2 minutes

2.4 Decision Support

Provide data support for engineering decisions.

Support Types:

Decision Type	Support Content
Construction Decision	Suggest construction pace based on monitoring data
Emergency Decision	Suggest emergency measures based on warning information
Maintenance Decision	Suggest maintenance plans based on long-term trends
Investment Decision	Suggest investment plans based on risk assessment

3. Usage Methods

3.1 WorkWise AI Assistant

1. Open WorkWise
2. Enter the “AI Assistant” page
3. Select AI model (local/cloud)
4. Enter natural language query
5. View AI response

3.2 CLI AI Assistant

# Query data
railwise ai query "Show the settlement trend of point JC-01"

# Generate report
railwise ai report --project "proj-001" --type weekly

# Analyze anomaly
railwise ai analyze --project "proj-001" --method anomaly

3.3 MCP Integration

Integrate AI Assistant through MCP Server into Claude, Cursor, and other AI clients.

4. Prompt Techniques

4.1 Effective Prompts

# Good Prompt
"Analyze the settlement data of points JC-01 to JC-10, find points with change rate exceeding 2mm/d, and explain possible causes and suggestions"

# Bad Prompt
"Analyze data"

4.2 Prompt Template

"As a monitoring engineer, analyze the [indicator] data of [point range], focus on [focus point], and provide [output format]"

5. Precautions

5.1 Data Privacy

• Local model: All data processing is done locally, data is not uploaded
• Cloud model: Data is uploaded to the cloud for processing, please confirm data sensitivity

5.2 Result Verification

• AI analysis results are for reference only and cannot replace professional engineer judgment
• Important decisions must be combined with actual site conditions

5.3 Model Selection

Model	Advantages	Disadvantages	Applicable Scenarios
Local	Data security, no network required	Limited capability	Data-sensitive projects
Cloud	Strong capability, intelligent	Data upload required	Complex analysis scenarios

6. Related Documents

• WorkWise AI Assistant
• AI Measurement Data Analysis
• AI Report Generation
• AI Anomaly Detection
• MCP Server Overview

---

RailWise · Smart Monitoring — Engineering Surveying and Safety Monitoring Digital Solutions