Publicação
Augmented AI for Decision Making: Leveraging Generative AI for Enhanced Decision-Making
| Resumo: | For more than a century, successive technological waves have reshaped managerial work, from Taylor’s stopwatch to enterprise resource-planning systems and, today, large language models. This research investigates the integration of Generative AI into the Business Process Management life cycle by positioning ChatGPT at the decision nodes of three different scenarios: strategic market entry, tactical budget allocation and operational staff scheduling. Therefore, a design-science approach combined with a critical literature review and an online survey captured 119 decisions from senior, middle and junior managers paired each with ChatGPT’s response to the same business case. Consistency, perceived feasibility, trust and willingness to delegate were analyzed with nonparametric statistics. Findings reveal a clear structure-ambiguity gradient: the model reliably cut planning time and achieved 86 % delegation acceptance in the highly structured operational task, matching managers in the tactical case only when its output was reviewed for soft constraints, and lost credibility at the strategic tier where political nuance dominated. The artifact of the study sets on analyzing both strengths and limitations on applying generative AI in business decision-making. This study delivers actionable guidance for responsible AI deployment: automate where rules dominate; retain human oversight where context, ethics, or stakeholder politics prevail; and embed explainability and governance at every stage. |
|---|---|
| Autores principais: | Sousa, Mário Henrique Toste |
| Assunto: | Artificial Intelligence Generative AI AI and data-driven decision-making Decision Making AI Limitations SDG 9 - Industry, innovation and infrastructure SDG 17 - Partnerships for the goals |
| Ano: | 2025 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso embargado |
| Instituição associada: | Universidade Nova de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório Institucional da UNL |
| Resumo: | For more than a century, successive technological waves have reshaped managerial work, from Taylor’s stopwatch to enterprise resource-planning systems and, today, large language models. This research investigates the integration of Generative AI into the Business Process Management life cycle by positioning ChatGPT at the decision nodes of three different scenarios: strategic market entry, tactical budget allocation and operational staff scheduling. Therefore, a design-science approach combined with a critical literature review and an online survey captured 119 decisions from senior, middle and junior managers paired each with ChatGPT’s response to the same business case. Consistency, perceived feasibility, trust and willingness to delegate were analyzed with nonparametric statistics. Findings reveal a clear structure-ambiguity gradient: the model reliably cut planning time and achieved 86 % delegation acceptance in the highly structured operational task, matching managers in the tactical case only when its output was reviewed for soft constraints, and lost credibility at the strategic tier where political nuance dominated. The artifact of the study sets on analyzing both strengths and limitations on applying generative AI in business decision-making. This study delivers actionable guidance for responsible AI deployment: automate where rules dominate; retain human oversight where context, ethics, or stakeholder politics prevail; and embed explainability and governance at every stage. |
|---|