Skip to main content

OpenAI Integration

AI-powered intelligence for engagement generation, order management, and worker creation.

Modules Provided​

Engagement Builder Module​

Capabilities:

  • Generate engagements from natural language
  • Parse order requests
  • Extract structured data
  • Validate completeness

Use Cases:

  • Email-to-order conversion
  • Voice-to-order processing
  • Chat-based ordering
  • Auto-quoting

OMS Intelligence Module​

Capabilities:

  • Order intent classification
  • Priority scoring
  • Issue detection
  • Recommendation generation

Use Cases:

  • Order routing
  • Priority assignment
  • Anomaly detection
  • Smart suggestions

Worker Builder Module​

Capabilities:

  • Generate worker code
  • Template selection
  • Code validation
  • Worker documentation

Use Cases:

  • Rapid worker prototyping
  • Worker scaffolding
  • Code generation
  • Development automation

Configuration​

OPENAI_API_KEY=<from-secrets>
OPENAI_MODEL=gpt-4
OPENAI_TEMPERATURE=0.1
OPENAI_MAX_TOKENS=2000

Bridge Access​

// Generate engagement from text
const engagement = await bridge.integrations.openAi.engagementBuilder
  .generateFromText({
    text: "I need 100 units of SKU-123 delivered to Chicago",
    context: { customerId, tenantId }
  });

// OMS intelligence
const analysis = await bridge.integrations.openAi.oms.analyzeOrder({
  engagement,
  history: customerOrders
});

// Generate worker
const workerCode = await bridge.integrations.openAi.workerBuilder
  .generateWorker({
    description: "Process inventory updates",
    template: "data-processor"
  });

Use Cases​

Natural Language Ordering: Convert customer messages into structured orders.

Intelligent Routing: Classify and route orders based on content analysis.

Development Acceleration: Generate worker boilerplate from descriptions.

Data Extraction: Parse unstructured data into engagement fields.

Advantages​

Intelligence: Understand context and intent, not just keywords.

Flexibility: Handle variations and unexpected inputs gracefully.

Speed: Process requests in seconds, not minutes.

Accuracy: High-quality structured output from unstructured input.

Operational Notes​

Cost Management​

  • Cache common requests
  • Use appropriate model for task
  • Set token limits
  • Monitor API usage

Reliability​

  • Implement retries
  • Handle rate limits
  • Fallback to manual processing
  • Validate AI output

Security​

  • No sensitive data in prompts
  • Sanitize inputs
  • Validate outputs
  • Audit AI decisions

When to Use​

Unstructured Input: When you need to parse natural language into structured data.

Intelligence Layer: When rule-based systems aren't flexible enough.

Development Tools: When you want to accelerate development with code generation.

When NOT to Use​

Deterministic Logic: Use regular code for predictable, deterministic operations.

High-Volume: AI calls have cost - use for value-add operations only.

Real-Time Critical: AI responses have latency - not for sub-second requirements.

OpenAI: Intelligence for commerce automation.