Artificial intelligence in digital marketing: methodology, advantages, and constraints

Methods. This study tests the pragmatic proposition that artificial intelligence (AI) works best as an effort multiplier, rather than a full-fledged replacement for human judgment. To this end, it traces how marketers integrate algorithms into strategy, content creation, and ongoing optimization, while addressing bias, data drift, and maintaining brand integrity. Methodologically, a consistent mixed-method design was applied. First, a systematic review of 112 peer-reviewed articles (2019–2025) established the theoretical basis for AI’s effectiveness and documented ethical and transparency risks. Second, a field study of 28 European SMEs, supported by a controlled twoweek A/B test in Meta Ads (≈ 240,000 impressions), quantified the real-world benefits. Key metrics – strategy preparation time, cost per lead (CPL), and perceived trust – were tracked, and partial least squares structural equation modeling (PLS-SEM) disentangled direct, indirect, and moderating effects. Results: Teams that used generative models in the ideation and copywriting stages reduced planning delays by 51%, but only achieved a 27% reduction in CPL when editors performed light proofreading; full automation further reduced editing time but doubled complaints about tone mismatch, increasing media spend. Personalization acted as a partial mediator, increasing CTR by 38% when the quality of the own data was above the 75th percentile. Trust in moderators affected cost effectiveness: the benefits disappeared when marketers expressed low confidence in machine outputs. Novelty. The paper introduces the concept of «curated acceleration» – an integrative framework that links each strategic phase (idea development, production, deployment) to achievable AI benefits and structural constraints (data quality, brand voice). This shifts the discourse from tool catalogs to workflow architecture and explains how human oversight and data hygiene together unlock the multiplier effect of AI. Practical value. AI provides tangible savings and more precise targeting only under the condition of strict data «hygiene» and minimal but conscious human oversight – what the paper calls «curated acceleration». Theoretically, an integrative framework is proposed that aligns strategic stages, achievable benefits and structural constraints, moving the discussion from a list of tools to the design of workflows. For practitioners, a decision map is provided that suggests where to increase prompts, where to slow down for editorial review, and when to output transparency signals. In short, AI scales creativity without sacrificing judgment, provided disciplined data practices and continuous monitoring are implemented.