Artificial intelligence is becoming part of everyday work in small and medium-sized enterprises, but using AI is not the same as building sophisticated AI-driven marketing analytics. A new multimethod study suggests that this distinction matters. Among 103 surveyed German SME representatives, 46% said their companies already used AI, yet the practical use of the technology was concentrated heavily in accessible tools and routine tasks rather than the more advanced analytical applications highlighted by the research literature.
The study, published in the Journal of Marketing Analytics on 3 October 2026, combined a systematic review of 49 studies with an exploratory survey of SMEs. Researchers Anita Talitha Parsegyan, Manuel Muth and Michael Lingenfelder used the two evidence streams to examine both what AI could potentially do in marketing analytics and the conditions that shape whether smaller firms can realistically adopt it.
A gap between AI availability and analytical depth
AI can support marketing research in ways that extend far beyond generating copy. The systematic review identified three broad application areas: synthetic data generation, text and content analysis, and AI-driven prediction. These applications can potentially help firms generate research data, analyse large volumes of customer language, identify patterns, forecast outcomes and support segmentation or other strategic decisions.
For SMEs, however, the ability to use these methods depends on more than whether an AI product exists. Smaller firms can face tighter financial constraints, less specialised expertise, weaker data infrastructure and fewer dedicated market-research resources than large organisations. The researchers therefore examined adoption from organisational, technological and user-related perspectives rather than treating AI adoption as a single yes-or-no decision.
How the researchers built the evidence base
The literature component followed PRISMA reporting principles and used Web of Science searches conducted in late October 2025. One search stream focused on AI adoption in SMEs, while a second focused on AI applications in marketing analytics and market research.
The SME adoption search initially returned 322 records. After filtering, screening and full-text assessment, 16 studies were included. The marketing-analytics search initially returned 1,058 records and ultimately contributed 33 studies. Together, the two streams produced a 49-study review. The researchers grouped recurring findings into organisational, technological and user-related adoption conditions, alongside the three main marketing-analytics application fields.
The empirical component was an exploratory status-quo survey conducted in 2024 with support from three chambers of industry and commerce. It included 103 participants from SMEs, mainly in Hesse, Germany. Decision-makers were strongly represented: 68% of respondents held top-level executive roles, 20% were in upper management and 12% were staff members. The survey was analysed descriptively rather than as a hypothesis-testing or causal study.
Nearly half already reported using AI
At the time of the survey, 46% of respondents said AI was already being used in their SME. Among the 54% whose firms were not using AI, 39% said implementation was planned within the following 12 months. When current use and planned adoption were considered together, the researchers calculated that 67% of the full sample was already seriously considering AI implementation in the organisation.
The perceived benefits were practical. Among SMEs already using AI, 83% identified time savings as a benefit and 51% identified cost savings. ChatGPT was the most commonly used form of AI, and reported tasks frequently involved text or image generation and administrative activities such as processing invoices, documents or customer enquiries.
Marketing was particularly prominent. Among participants whose SMEs already used AI, 53% reported using it in marketing and 32% in sales. The authors nevertheless found that market research, sales controlling and category management were among the less frequently reported application areas. This is important because the literature review identified considerably more advanced possibilities for AI-enabled marketing analytics than were visible in the surveyed firms’ current practices.
Legal uncertainty and data problems remained visible
The survey also showed that adoption does not eliminate uncertainty. Among AI users, 49% reported unclear legal regulations as a challenge, compared with 32% of non-users. In addition, 43% of both AI users and firms planning implementation wanted assistance with legal and ethical questions.
Data presented another practical barrier. About one third of both users and non-users identified the effort required to ensure data availability and quality as a challenge. Data security was cited by 32% of AI users, while 36% of non-users reported uncertainty about data-protection compliance. A lack of transparency in AI results was reported by 34% of users and 23% of non-users, and 30% of users identified employee errors in the use of AI as an issue.
Training, by contrast, did not emerge simply as an unwanted burden. Forty-three percent of all participants expressed interest in additional training for employees and managers, rising to 73% among those planning AI implementation. Only 2% of existing AI users identified lack of expertise as a challenge. The authors also noted that employee acceptance remained relevant, with lack of acceptance reported by 28% of organisations already implementing AI and 18% of organisations without AI.
Four ways SMEs may approach AI
Bringing the review and survey together, the researchers proposed a preliminary framework based on two dimensions: the favourability of AI adoption conditions and the amount of useful AI application potential available to the SME. This produces four configurations described as opportunistic users, operational users, strategically ready firms and strategic users.
Opportunistic users have relatively unfavourable adoption conditions and limited application potential, so AI may remain a selective tool used by individuals for particular efficiency gains. Operational users may have substantial application potential in a specific area while lacking broad organisational conditions or governance for strategic adoption. Strategically ready firms have favourable organisational, technological and user conditions but comparatively limited relevant applications. Strategic users combine favourable adoption conditions with broad application potential.
The authors do not present these four configurations as a ladder in which every business should aim for maximum AI use. The appropriate level depends on the firm’s objectives and resources. That qualification is particularly relevant for SMEs because an inexpensive, narrowly targeted AI tool may create genuine value without requiring the organisation to reproduce the infrastructure of a large corporation.
What the findings mean for smaller businesses
The study suggests that headline measures of business AI adoption can hide substantial differences in what companies are actually doing. A firm using generative AI for copy drafting and routine administration is an AI user, but its analytical capabilities are very different from a firm using machine learning for customer prediction, automated text analysis or synthetic research data.
For managers, the practical implication is that adoption decisions should begin with a defined problem rather than with AI as an objective in itself. Data quality, compatibility with existing systems, management support, employee acceptance, usability, governance and legal clarity can determine whether a technically promising application becomes useful in practice. The survey’s strong emphasis on time savings also suggests that many SMEs currently experience AI first as an efficiency technology rather than as a fully developed source of strategic customer insight.
The discrepancy between literature and practice may also represent an opportunity. The review found possibilities for synthetic data, text analysis and predictive methods that remain comparatively underused in the surveyed firms. As tools become easier to deploy, some capabilities that previously required specialist teams may become more accessible to smaller businesses. Whether this produces better marketing decisions, however, requires direct empirical testing rather than assumption.
Important limitations
The study is exploratory and should not be interpreted as a representative estimate of AI use among all German or European SMEs. The survey included only 103 respondents and was geographically concentrated mainly in Hesse. Its descriptive design does not establish that any reported adoption condition causes firms to implement AI or obtain better results.
Timing is another important limitation in a fast-moving field. The survey was conducted in 2024, while the literature search was conducted in 2025. AI tools and organisational practices can change rapidly, meaning that a paper published in 2026 necessarily describes evidence collected at earlier stages of the technology’s diffusion. The authors recommend repeated surveys, broader geographic sampling and future deductive research capable of testing causal relationships.
The proposed four-part framework is also conceptual. It integrates patterns from the literature and survey, but it has not yet been empirically validated as a typology, and the adoption-condition dimension has not yet been quantitatively operationalised. It should therefore be treated as a structured way to think about SME AI use rather than a proven classification system.
A more precise view of business AI adoption
The research offers a useful distinction at a time when AI adoption statistics can make business transformation appear more uniform than it is. In this SME sample, AI was already common enough to affect many business functions, especially marketing, yet much of that use centred on accessible tools and immediate efficiency tasks.
The next stage of adoption may therefore depend less on whether SMEs use AI at all and more on what they use it for, whether their data and governance can support it, and whether advanced analytical applications solve problems worth the investment. For smaller firms, the most valuable AI strategy may not be the broadest one, but the one that matches a credible use case to the organisation’s actual capabilities.
Source Information
Study: AI-enabled marketing analytics for SMEs: multimethod evidence on application potential and adoption conditions
Authors: Anita Talitha Parsegyan, Manuel Muth and Michael Lingenfelder
Journal: Journal of Marketing Analytics
Published: 3 October 2026
DOI: 10.1057/s41270-026-00539-2
