Publication Details
Abstract
Small and medium-sized enterprises (SMEs) are rapidly adopting artificial intelligence (AI) tools in marketing, yet evidence on how these tools change marketing operating costs remains fragmented and dominated by large-firm cases. This paper develops a transparent, activity-based cost model of monthly marketing operating expenditure (OPEX) and evaluates it through a Monte Carlo experiment with 10,000 draws per scenario. Three scenarios are compared for a representative SME marketing unit: a manual baseline (S0), partial AI adoption through off-the-shelf software-as-a-service tools (S1), and integrated AI adoption built on a CRM and marketing-automation core (S2). Costs comprise labour, agency spending, media-spend wastage, tooling, amortised implementation and training, and AI oversight. For a three-person team, mean monthly OPEX falls from USD 8,879 (S0) to USD 8,362 under S1 (−5.8%) and to USD 8,523 under S2 (−3.9%). Although S2 removes far more gross cost (USD 2,150 vs. 1,135 per month), it also adds USD 1,795 of AI-related cost, so its net advantage is smaller and far less certain (probability of positive savings 79.0% vs. 99.1% for S1). Scale is decisive: integrated AI does not pay off for one- or two-person teams, and it overtakes partial AI only from roughly four to five marketing employees. Sensitivity analysis shows that the share of freed labour hours actually converted into cost savings and the recurring licence burden dominate outcomes. The findings support a staged, measurement-led adoption strategy.