Fatores críticos de sucesso no uso de agentes com inteligência artificial no atendimento

Resumo

The rapid adoption of artificial intelligence virtual agents in customer service has heightened expectations for efficiency and personalization. Nevertheless, the failure rate of such initiatives remains substantial. This study investigates the critical success factors (CSFs) that shape the performance of these agents and proposes a hierarchical taxonomy to guide the implementation of AI-powered virtual agents in the post–Large Language Model (LLM) era. Its primary objective is to provide a reference framework to support the prioritization of organizational and technological capabilities. The research adopts a multi-stage methodological design aimed at theory building and the assessment of robustness and significance: (1) construction of the bibliographic sample through a systematic protocol of search, screening, and selection of recent evidence (2023–2025), encompassing 160 academic articles and technical reports; (2) reflexive thematic analysis and evidence coding; (3) binary metaaggregation and incidence quantification; (4) classification and hierarchical structuring based on threshold criteria, distinguishing different levels of universality; (5) stability testing through deterministic sensitivity analysis and Monte Carlo simulation; and (6) external validation of the taxonomy by four domain specialists. The findings consolidate 27 factors organized into three levels of criticality, comprising 5 Universal pillars (Reliability, Quality, Technical Competence, User Experience, and Humanization), 2 Moderators (Response Fluency and Journey Support), and 20 Emerging factors, in addition to contingent factors and sector-specific critical factors across six sectors of the economy. The study offers a robust reference framework that may support the planning, implementation, and management of AI-enabled customer service initiatives.

Ficha do documento

Tipo

Dissertação

Ano

2026

Instituição

Fundação Getulio Vargas

Fonte

Repositório da FGV

Idioma

Português

Acesso

Acesso aberto

Identificador

oai:repositorio.fgv.br:10438/38505

Temas

Tecnologia

Palavras-chave

ChatbotsCritical success factorsGenerative artificial intelligenceCustomer serviceVirtual agentsAgentic AIIT-Information TechnologyFatores críticos de sucessoInteligência artificial generativaAtendimento ao clienteAgentes virtuaisIA agênticaTI-Tecnologia de InformaçãoAdministração de empresasServiços ao cliente - Inovações tecnológicasInteligência artificialInteração homem-máquinaSucesso nos negócios