Hasan, Md Majedul and Mohammad, Abdul Rahman and Bhuiyan, Mohd Salman Hossain and Sukhi, Rabia Akter (2026) Closing the AI-to-Action Gap: A Human-Centered Explainable AI Framework for Marketing Resource Allocation in Resource-Constrained U.S. Small Businesses. American Journal of Technology Advancement, 3 (10). ISSN 2997-9382

Text AJTA_Closing+the+AI-to-Action (1).pdf Download (1MB)

Abstract

While small businesses are increasingly turning to digital marketing to attract customers, make money and stay competitive, the allocation of marketing resources is often difficult to achieve due to limited financial and operational resources. While AI can provide predictions about campaign success and opportunities for high-value marketing activities, many small businesses may find it challenging to make sense of AI-driven insights, build trust in the data, and act on it. This study aims to present a human-centric Explainable Artificial Intelligence (XAI) framework to bridge the gap between AI and action in marketing resource allocation for resource scarcity situations of U.S. small businesses. The data set includes 10,000 observations of each advertising campaign and 41 variables related to the advertising platform, campaign goal, audience, creative, budget level, engagement, ad spend, conversions, revenue and profitability. Various machine learning models are utilized such as Multiple Linear Regression, Random Forest, XGBoost, and LightGBM for predicting the profitability of the campaign and the factors related to the effectiveness of the marketing performance. SHAPley Additive explanations (SHAP) are built-in to offer transparent interpretations of model predictions and explain what factors make up the basis for AI-powered marketing suggestions. The proposed framework builds on these predictions and explanations to enable budget allocation decisions on marketing platforms and campaign configurations with respect to resource constraints. The framework highlights the importance of human decision making by providing AI recommendations in the form of explainable recommendations that allow managers to understand, evaluate, accept, adjust, and reject proposed allocations. Predictive accuracy and error measures are used to measure model performance, and profitability, return on advertising investment and resource efficiency are used to measure the effectiveness of the allocation. It is the inclusion of the decision-oriented approach, which extends the traditional marketing prediction method by combining predictive analytics, explainability, resource optimization, and human judgment that sets the study apart. The proposed framework offers a tangible strategy for turning marketing insights derived from AI into resource allocation decisions, and can overcome issues of transparency, usage and resource constraints for small businesses.

Item Type:	Article
Uncontrolled Keywords:	Explainable Artificial Intelligence (XAI), Marketing Resource Allocation, Human-Centered AI, Small Businesses, Marketing Profitability and Artificial Intelligence Decision Support
Subjects:	H Social Sciences
Depositing User:	admin eprints
Date Deposited:	04 Nov 2026 07:26
Last Modified:	04 Nov 2026 07:26
URI:	http://eprints.umsida.ac.id/id/eprint/17412

Actions (login required)

View Item