Multi-Agent Vendas Python

Sistema multi-agente para automacao do processo comercial: 5 agentes IA especializados que operam em sequencia, do diagnostico ao fechamento.

Agentes

Agente	Missao
diagnostico.py	Identifica o maior problema do negocio via perguntas abertas
qualificador.py	Avaliacao BANT (Budget, Authority, Need, Timeline) com score 0-100
analise_call.py	Extrai insights de transcricoes de reunioes
orcamento.py	Gera proposta comercial personalizada baseada no diagnostico
simulador.py	Treino de objecoes — prospect IA + coach que avalia suas respostas

Fluxo

Lead novo
    |
diagnostico.py
(3-5 perguntas)
    |
qualificador.py
(BANT score)
    |
score >= 60? ──NAO──> nutrir / desqualificar
    |
   SIM
    |
orcamento.py
(proposta JSON)
    |
analise_call.py ←── transcricao da reuniao
    |
simulador.py ←── treino pre-call

Instalacao

git clone https://github.com/Dimitrearaujo/multi-agent-vendas-python
cd multi-agent-vendas-python
pip install -r requirements.txt

cp .env.example .env
# Edite .env com sua chave Anthropic

Uso

# Fluxo completo interativo: diagnostico + BANT + proposta
python orchestrator.py diagnostico

# Analisar transcricao de reuniao
python orchestrator.py call reuniao.txt

# Treino de objecoes
python orchestrator.py simulador "automacao IA para clinicas"

Uso dos agentes individualmente

from agents import diagnostico, qualificador, orcamento

# Diagnostico
historico = [{"role": "user", "content": "Tenho uma clinica vet com 3 funcionarios"}]
resultado = diagnostico.run(historico)
# {"pergunta": "Qual e o maior problema hoje?", "diagnostico": null}

# Qualificacao BANT
bant = qualificador.qualificar(historico)
# {"score": 75, "recomendacao": "qualificado", ...}

# Gerar proposta
proposta = orcamento.gerar(diagnostico_resultado, {"nome": "Dr. Silva"})

Estrutura

multi-agent-vendas-python/
   agents/
      diagnostico.py     # Perguntas + sintese do problema
      qualificador.py    # Score BANT
      analise_call.py    # Insights de reunioes
      orcamento.py       # Proposta comercial JSON
      simulador.py       # Treino de objecoes
   core/
      llm.py             # Wrapper Claude API
   orchestrator.py       # Fluxo interativo completo

Licenca

MIT

---

🇺🇸 English

Multi-Agent Sales Python

Multi-agent system for automating the sales process: 5 specialized AI agents operating in sequence, from diagnosis to closing.

Agents

Agent	Mission
diagnostico.py	Identifies the biggest business problem via open-ended questions
qualificador.py	BANT evaluation (Budget, Authority, Need, Timeline) with 0-100 score
analise_call.py	Extracts insights from meeting transcripts
orcamento.py	Generates a personalized commercial proposal based on diagnosis
simulador.py	Objection training — AI prospect + coach that evaluates your responses

Flow

New lead
    |
diagnostico.py
(3-5 questions)
    |
qualificador.py
(BANT score)
    |
score >= 60? ──NO──> nurture / disqualify
    |
  YES
    |
orcamento.py
(JSON proposal)
    |
analise_call.py <── meeting transcript
    |
simulador.py <── pre-call training

Installation

git clone https://github.com/Dimitrearaujo/multi-agent-vendas-python
cd multi-agent-vendas-python
pip install -r requirements.txt

cp .env.example .env
# Edit .env with your Anthropic key

Usage

# Full interactive flow: diagnosis + BANT + proposal
python orchestrator.py diagnostico

# Analyze meeting transcript
python orchestrator.py call meeting.txt

# Objection training
python orchestrator.py simulador "AI automation for clinics"

Individual agent usage

from agents import diagnostico, qualificador, orcamento

# Diagnosis
history = [{"role": "user", "content": "I have a vet clinic with 3 employees"}]
result = diagnostico.run(history)
# {"question": "What's the biggest problem today?", "diagnosis": null}

# BANT qualification
bant = qualificador.qualificar(history)
# {"score": 75, "recommendation": "qualified", ...}

# Generate proposal
proposal = orcamento.gerar(diagnosis_result, {"name": "Dr. Silva"})

Structure

multi-agent-vendas-python/
   agents/
      diagnostico.py     # Questions + problem synthesis
      qualificador.py    # BANT score
      analise_call.py    # Meeting insights
      orcamento.py       # Commercial proposal JSON
      simulador.py       # Objection training
   core/
      llm.py             # Claude API wrapper
   orchestrator.py       # Full interactive flow

License

MIT

---

← Back to profile