Skip to main content

AI Classifier

Type: ai_classifier • Category: flow • Tags: ai, classifier, decision, routing, llm

Description

AI-powered decision maker for workflow routing

Parameters

NameTypeDescriptionRequiredDefault
promptstringInstructions for AI on how to classify the data (include classification options like 'Classify as: a, b, c')no
dataExpstringno

Help

AI Classifier Worker

Uses AI to analyze data and make classification decisions for workflow routing. Perfect for intelligent decision-making based on complex data patterns.

How it works:

  1. Provide classification instructions in the prompt
  2. Specify data expression to extract data for analysis
  3. AI returns a classification decision

Financial Markets Examples:

  • Stock Classification: "Classify stock data as: gold, nasdaq, crypto, forex" - analyzes stock data to determine market type
  • Market Sentiment: "Analyze news sentiment: bullish, bearish, neutral" - classifies market news sentiment
  • Trading Signals: "Generate signal: buy, sell, hold based on technical indicators" - creates trading decisions from indicators
  • Risk Assessment: "Assess risk level: low, medium, high based on volatility and volume" - evaluates investment risk
  • Market Regime: "Classify market regime: trending, ranging, volatile, calm" - determines current market conditions
  • Asset Allocation: "Recommend allocation: stocks, bonds, cash, commodities" - suggests portfolio allocation
  • News Impact: "Classify news impact: major, moderate, minor" - assesses importance of financial news

General Examples:

  • Route customer inquiries: "Classify as: support, sales, billing, technical"
  • Content analysis: "Classify content type: news, opinion, analysis, spam"
  • Decision making: "Decide: approve, reject, review"

Decision Output: Returns {"decision": "classification_result"}

Use this result in workflow conditions to route execution paths.

Tips:

  • Be specific in your prompt about classification criteria
  • Include possible options in your prompt (e.g., "Classify as: option_a, option_b, option_c")
  • For financial data, specify what aspects to analyze (price, volume, indicators, news)
  • Test with sample data to verify accuracy