Prompt Details
Model
Chat - (gpt-5.4)
Token size
240
Example input
[TASK]: Build a customer support chatbot that routes tickets to specialized agents for billing, technical issues, and refunds
Example output
1. AGENT ROSTER
Triage Router: Classify incoming ticket by intent and urgency. Input: customer message. Output: category, priority, extracted entities. Model: GPT-4o-mini. Tools: none. Failure: default to general support queue.
Billing Agent: Handle invoices, payment issues, subscription changes. Input: classified ticket plus customer account data. Output: resolution or escalation flag. Model: GPT-4o. Tools: Stripe API, CRM lookup. Failure: escalate to human with context summary.
Tech Support Agent: Diagnose technical problems, suggest fixes. Input: classified ticket plus product logs. Output: step-by-step resolution. Model: GPT-4o. Tools: knowledge base search, log parser. Failure: create engineering ticket automatically.
Refund Agent: Process refund requests per company policy. Input: classified ticket plus order history. Output: approved/denied with reason. Model: GPT-4o-mini. Tools: order database, refund policy rules engine. Failure: escalate to manager queue.
Quality Monitor: Review completed resolutions for tone and accuracy. Input: full conversation transcript. Output: quality score plus flags. Model: GPT-4o-mini. Tools: sentiment analyzer. Failure: log silently, do not block.
By purchasing this prompt, you agree to our terms of service
GPT-5.4
Design production-ready multi-agent LLM systems in minutes. Get a complete architecture: agent roster with I/O contracts and failure modes, orchestration pattern with justification, context protocol with JSON schemas, error handling with retries and circuit breakers, anti-patterns specific to your use case, and implementation skeleton. Works with CrewAI, LangGraph, AutoGen, Strands Agents, OpenAI Agents SDK, or framework-agnostic.
...more
Added over 1 month ago
