Paper Number
ICIS2025-2394
Paper Type
Short
Abstract
Digital transformation initiatives frequently fail due to infrastructure deployment bottlenecks, with up to 84% of projects falling short of expectations. Traditional manual provisioning methods create delays, misalignment with business goals, and high operational costs. This paper introduces Prompt2Infra, a generative multi-agent AI framework that translates natural language requirements into cloud-deployable infrastructure solutions. Using Design Science Research methodology, we developed a five-agent system (Specification, CodeGen, Validation, Deployment, and Monitoring) with embedded governance and recursive feedback mechanisms. Evaluation through 20 industry-inspired simulation scenarios demonstrates significant improvements: 45% reduction in provisioning time, 30% lower operational costs, and decreased deployment failure rates from 12% to 4.5%. The framework enables non-technical stakeholders to participate meaningfully in infrastructure specification while maintaining compliance and security standards. This work contributes three key design principles, role-specific prompting, compliance-as-code, and feedback-driven orchestration. This research advancing theoretical understanding of human-AI collaboration in digital transformation contexts.
Recommended Citation
Gohel, Hardik; Raval, Maulin; and Wan, Yun, "Human-AI Collaboration for Organizational Transformation through Generative Infrastructure Design" (2025). ICIS 2025 Proceedings. 26.
https://aisel.aisnet.org/icis2025/gen_ai/gen_ai/26
Human-AI Collaboration for Organizational Transformation through Generative Infrastructure Design
Digital transformation initiatives frequently fail due to infrastructure deployment bottlenecks, with up to 84% of projects falling short of expectations. Traditional manual provisioning methods create delays, misalignment with business goals, and high operational costs. This paper introduces Prompt2Infra, a generative multi-agent AI framework that translates natural language requirements into cloud-deployable infrastructure solutions. Using Design Science Research methodology, we developed a five-agent system (Specification, CodeGen, Validation, Deployment, and Monitoring) with embedded governance and recursive feedback mechanisms. Evaluation through 20 industry-inspired simulation scenarios demonstrates significant improvements: 45% reduction in provisioning time, 30% lower operational costs, and decreased deployment failure rates from 12% to 4.5%. The framework enables non-technical stakeholders to participate meaningfully in infrastructure specification while maintaining compliance and security standards. This work contributes three key design principles, role-specific prompting, compliance-as-code, and feedback-driven orchestration. This research advancing theoretical understanding of human-AI collaboration in digital transformation contexts.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
12-GenAI