Automation Decision Engine

Rule-based and event-driven decision engine for business automation—handling complex logic, routing, and execution without manual intervention.

---

🎯 Overview

This engine automates business decisions by evaluating rules, processing events, and executing actions—enabling consistent, repeatable execution at scale.

┌─────────────┐     ┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│   Event     │────▶│    Rule     │────▶│  Decision   │────▶│   Action    │
│   Input     │     │   Engine    │     │   Output    │     │   Executor  │
└─────────────┘     └─────────────┘     └─────────────┘     └─────────────┘
                          │
                          ▼
                    ┌─────────────┐
                    │   Audit     │
                    │    Log      │
                    └─────────────┘

---

✨ Features

• Rule Engine - Define complex business rules in YAML or code
• Decision Trees - Visual decision flow with branching logic
• Event Processing - React to real-time events from any source
• Action Execution - Trigger workflows, APIs, notifications
• Audit Trail - Complete logging of all decisions for compliance
• A/B Testing - Test different rule sets in production
• Fallback Handling - Graceful degradation when rules don't match

---

🏗️ Architecture

src/
├── api/                 # REST API
│   ├── decisions.py
│   ├── rules.py
│   └── events.py
├── core/                # Core engine
│   ├── rule_engine.py
│   ├── decision_tree.py
│   ├── evaluator.py
│   └── context.py
├── rules/               # Rule definitions
│   ├── parser.py
│   ├── validator.py
│   └── compiler.py
├── actions/             # Action executors
│   ├── base_action.py
│   ├── webhook_action.py
│   ├── email_action.py
│   └── workflow_action.py
├── storage/             # Persistence
│   ├── rule_store.py
│   └── decision_log.py
└── monitoring/          # Observability
    ├── metrics.py
    └── audit.py

---

🚀 Quick Start

# Clone repository
git clone https://github.com/daveedashar/automation-decision-engine.git
cd automation-decision-engine

# Install dependencies
pip install -r requirements.txt

# Run the service
python -m src.main

---

📋 Rule Definition (YAML)

# rules/lead_routing.yaml
name: lead_routing
version: 1.0
description: Route leads based on score and region

rules:
  - name: high_value_enterprise
    conditions:
      all:
        - field: lead_score
          operator: gte
          value: 80
        - field: company_size
          operator: eq
          value: "enterprise"
    actions:
      - type: assign
        to: "enterprise_team"
      - type: notify
        channel: "slack"
        message: "🔥 High-value lead: {{lead.name}}"
    priority: 100

  - name: mid_market_qualified
    conditions:
      all:
        - field: lead_score
          operator: gte
          value: 50
        - field: company_size
          operator: in
          value: ["mid_market", "smb"]
    actions:
      - type: assign
        to: "sales_team"
      - type: add_to_sequence
        sequence: "nurture_qualified"
    priority: 50

  - name: default_routing
    conditions:
      all: []  # Catch-all
    actions:
      - type: assign
        to: "sdr_team"
    priority: 0

---

📋 Usage Example

Evaluate a Decision

from src.core import DecisionEngine

engine = DecisionEngine()
engine.load_rules("rules/lead_routing.yaml")

# Evaluate
result = engine.evaluate({
    "lead_score": 85,
    "company_size": "enterprise",
    "region": "north_america",
    "lead": {
        "name": "Acme Corp",
        "email": "[email protected]"
    }
})

print(result)
# {
#     "rule_matched": "high_value_enterprise",
#     "actions": [
#         {"type": "assign", "to": "enterprise_team"},
#         {"type": "notify", "channel": "slack", ...}
#     ],
#     "decision_id": "dec_abc123",
#     "timestamp": "2026-01-03T22:30:00Z"
# }

Define Rules in Code

from src.core import Rule, Condition, Action

rule = Rule(
    name="fraud_detection",
    conditions=[
        Condition("transaction_amount", "gte", 10000),
        Condition("country", "in", ["high_risk_countries"]),
        Condition("velocity_24h", "gte", 5),
    ],
    match_type="all",  # or "any"
    actions=[
        Action("flag", severity="high"),
        Action("notify", team="fraud_ops"),
        Action("hold_transaction"),
    ]
)

engine.add_rule(rule)

---

⚙️ Operators

Operator	Description	Example
eq	Equals	field: status, operator: eq, value: "active"
neq	Not equals	field: type, operator: neq, value: "test"
gt	Greater than	field: amount, operator: gt, value: 100
gte	Greater than or equal	field: score, operator: gte, value: 50
lt	Less than	field: age, operator: lt, value: 30
lte	Less than or equal	field: count, operator: lte, value: 10
in	In list	field: country, operator: in, value: ["US", "UK"]
not_in	Not in list	field: status, operator: not_in, value: ["banned"]
contains	Contains substring	field: email, operator: contains, value: "@gmail"
regex	Regex match	field: phone, operator: regex, value: "^\+1"

---

📊 Decision Flow

EVENT RECEIVED
      │
      ▼
┌─────────────┐
│  Load Rules │
│  (Priority) │
└─────────────┘
      │
      ▼
┌─────────────┐     NO      ┌─────────────┐
│  Evaluate   │────────────▶│   Next      │
│  Conditions │             │   Rule      │
└─────────────┘             └─────────────┘
      │ YES                        │
      ▼                            │
┌─────────────┐                    │
│  Execute    │                    │
│  Actions    │                    │
└─────────────┘                    │
      │                            │
      ▼                            │
┌─────────────┐                    │
│  Log        │◀───────────────────┘
│  Decision   │     (if no match)
└─────────────┘

---

🧪 Testing

# Run tests
pytest tests/ -v

# Test specific rules
python -m src.cli test-rule rules/lead_routing.yaml --input test_data.json

---

📈 Outcomes

• Eliminated manual decision bottlenecks
• 100% consistent execution across all cases
• < 50ms decision latency
• Complete audit trail for compliance

---

📄 License

MIT License - see LICENSE for details.

---

👤 Author

Daud Ashar

• GitHub: @daveedashar
• LinkedIn: /in/daudashar
• Email: [email protected]