Voltar à galeria

Agentes de IA

Customer Support Agent Workflow

Define a bounded support-agent decision loop with tool use and human escalation.

Modelos
ChatGPT, Claude
Nível
Avançado
Versão
v0.9

Este prompt está disponível em inglês. A interface permanece no idioma selecionado.

01

O que este prompt faz

Define a bounded support-agent decision loop with tool use and human escalation.

02

Ideal para

JSON containing intent, facts, missing information, planned tool calls, policy checks, response, escalation status, and audit note.

03

Variáveis

[PRODUCT_POLICY][TOOLS][ESCALATION_RULES][CUSTOMER_MESSAGE]
Prompt do sistema
You design safe support workflows. Policies and tool permissions are authoritative; never promise an action that a tool did not confirm.
Prompt do usuário
OBJECTIVE:
Handle [CUSTOMER_MESSAGE] using [PRODUCT_POLICY], available tools [TOOLS], and [ESCALATION_RULES].

INPUTS:
PRODUCT_POLICY: [PRODUCT_POLICY]
TOOLS: [TOOLS]
ESCALATION_RULES: [ESCALATION_RULES]
CUSTOMER_MESSAGE: [CUSTOMER_MESSAGE]

OUTPUT CONTRACT:
JSON containing intent, facts, missing information, planned tool calls, policy checks, response, escalation status, and audit note.

Como usar

Replace every variable with specific, non-sensitive information. Remove irrelevant sections and add one representative example when tone or classification matters.

Por que funciona

The blueprint separates authority, task context, evidence limits, and the output contract so requirements can be checked independently.

Compatibilidade

ChatGPTuntestedEditorial compatibility hypothesis; cross-model test record pending.

ClaudeuntestedEditorial compatibility hypothesis; cross-model test record pending.

05

Entrada de exemplo

PRODUCT_POLICY: [PRODUCT_POLICY]
TOOLS: [TOOLS]
ESCALATION_RULES: [ESCALATION_RULES]
CUSTOMER_MESSAGE: [CUSTOMER_MESSAGE]

06

Saída de exemplo

A structured deliverable following the listed output contract, with assumptions and missing evidence clearly marked.

Como personalizar

Replace every variable with specific, non-sensitive information. Remove irrelevant sections and add one representative example when tone or classification matters.

Erros comuns

Do not paste secrets or personal data. Avoid contradictory constraints, unsupported facts, and using generated output without review.

Notas de teste

Editorial blueprint awaiting recorded cross-model evaluation. It is visible for review but is not labeled as tested.

Why is this blueprint not marked as tested?

It has passed editorial structure review but does not yet have a published cross-model test record.