JISARA

Abstract: As artificial intelligence (AI) models become more prevalent across all fields, streamlined development of these models is becoming increasingly necessary. Deploying an AI model consists of integration with existing systems, monitoring various metrics related to the model, and maintenance of the model to keep it functional and up to date. Thus, successful deployment ensures value and sustainability. The objective of this research is to (1) identify the best practices within the phases of deployment, (2) explore cost requirements as well as strategies for savings, and (3) identify priorities during deployment. To explore these objectives a semi-structured interview paradigm was developed and conducted with AI professionals with a qualitative analysis performed on the resulting transcripts. The analysis showed that participants emphasized explainable models that were accessible to users. Deployment costs were highly dependent on where the model was hosted and whether the model was developed in house or acquired from the commercial market. Finally, priorities were dependent on the type of model being developed, the users it would interact with, and the data it was handling. Regardless of these factors, all participants highlighted the importance of explainability, accessibility, and cost. These factors were prioritized by participants during model deployment.

Download this article: JISARA - V19 N3 Page 17.pdf

Recommended Citation: Williams, N., Cummings, J., Wang, Y., Gokce, Y., (2026). AI & Machine Learning Deployment: Best Practices, Costs and Priorities. Journal of Information Systems Applied Research and Analytics 19(3) pp 17-27. https://doi.org/10.62273/ZVNR9663