Two SMEs can have access to the same AI tools and still differ substantially in their ability to use them effectively. That is why AI readiness cannot be reduced to technology access alone.

This article is informed by my ongoing doctoral research into AI readiness in SMEs in Mauritius. The research is focused on developing and validating a theory-anchored readiness instrument. The discussion below is therefore research-informed and conceptual; it should not be read as a report of completed empirical findings.

Why readiness is multidimensional

Organisational-adoption research has consistently treated technology adoption as shaped by more than the technology itself. The Technology–Organization–Environment (TOE) framework (Tornatzky & Fleischer, 1990) is the foundational model here, distinguishing the technology available to a firm, the organisation adopting it, and the external environment it operates in as three separate contexts shaping adoption outcomes. More recent AI-readiness research has extended this further: Jöhnk, Weißert and Wyrtki's 2021 interview study identifies organisational AI readiness as spanning strategic alignment, resources, knowledge, culture, and data, categories that go well beyond simply owning the right software.

A specific extension relevant here is TOE-H (Technology, Organisation, Environment, and Human), which adds an explicit human dimension to the original three. This isn't a term I've coined: Naheed, Pinto and Pirola (2025), writing in Procedia Computer Science (vol. 253, pp. 774–783), propose a preliminary multidimensional AI-readiness assessment model for SMEs built on exactly this TOE-H structure, including an expert-review-based assessment methodology. My doctoral research draws on the same four contextual dimensions while developing and empirically validating an instrument for the Mauritian SME context. How that instrument differs from, extends, or improves upon existing readiness models is part of the contribution that the research itself must establish.

The TOE-H lens

Applied to SME AI readiness, the four dimensions ask different questions:

Technology

Core question: is the technology itself accessible and fit for purpose?
  • Readiness looks like: usable, affordable tools matched to a named problem
  • Failure looks like: buying capability the business has no specific use for

Organisation

Core question: does the organisation have the structure to use it responsibly?
  • Readiness looks like: clear data ownership, a named use-case owner, basic governance
  • Failure looks like: a capable tool with no one accountable for its outputs

Environment

Core question: what external conditions shape the decision to adopt?
  • Readiness looks like: a reasonably clear sense of the applicable regulatory and market context
  • Failure looks like: adopting (or avoiding) AI based on rumour about what the law requires

Human

Core question: are the people who'll use and be affected by the system actually ready?
  • Readiness looks like: staff able to question outputs rather than accept or reject them reflexively
  • Failure looks like: a tool introduced with no attention to trust, skill, or workflow disruption

The TOE-H AI readiness lens: a conceptual tool, not a scored or validated instrument.

Technology: necessary, rarely sufficient

Technology readiness concerns whether suitable tools and supporting infrastructure are available for the business problem in question. Tool availability alone, however, says little about whether the organisation has the data, ownership, governance or human capability needed to use those tools effectively. This is the dimension most public discussion of AI adoption focuses on, and the one that, on its own, explains the least about whether adoption actually succeeds.

Organisation: the quiet determinant

Data ownership, use-case clarity, and basic governance sit in this dimension. An organisation can have excellent tool access and still be unready here: no one accountable for an AI-assisted output, no documented rule about what data can be used where. This is consistent with the broader AI-readiness literature's emphasis on organisational conditions as distinct from, and often more binding than, technology availability.

Environment: policy and market context

External conditions (national AI policy, data protection law, competitive pressure) shape adoption decisions, sometimes through accurate understanding and sometimes through rumour. In a small-island context, it is worth examining whether factors such as market scale, talent availability, institutional support and regulatory context alter the relative importance of readiness dimensions. This article does not assume that they do; that requires evidence. In Mauritius specifically, the actual, current state of relevant policy is set out in what the National AI Strategy actually says for SME owners, including where genuine legal obligations already exist under the Data Protection Act 2017, independent of newer, still-developing AI-specific policy.

Human: the dimension hardest to measure and easiest to underweight

Leadership mindset, staff trust in algorithmic outputs, and interpretive capacity all plausibly shape whether adoption actually sticks once a tool is introduced, a proposition consistent with the wider organisational-readiness literature's emphasis on human and cultural factors, not a claim specific to any dataset from this research. A live, unanswered question (one this doctoral research is designed to help investigate, not one it has already answered) is: how do employee trust, perceived job threat, skills, workflow disruption, and managerial implementation approach affect AI adoption at the human level? Framing a tool purely as an efficiency or headcount measure, versus a capability-augmentation one, is a plausible factor in how willingly staff engage with it, but the specific mechanisms in a Mauritian SME context remain to be tested, not assumed.

Questions an SME should ask before claiming readiness

  • Technology: is this tool matched to a specific, named problem, or acquired because it's available?
  • Organisation: who is accountable for this tool's outputs, and is that written down anywhere?
  • Environment: do we actually know what current law and policy require, or are we assuming?
  • Human: have we asked staff what they're worried about, or just told them what to use?

An organisation that can't answer one or more of these clearly has identified, specifically, where its readiness gap sits, which is more useful than a single overall "are we ready" verdict.

What the doctoral research is trying to establish

My doctoral research is developing and validating a theory-anchored AI readiness instrument for SMEs in Mauritius using a staged research design that includes conceptualisation, qualitative item generation, expert review, pilot testing with exploratory factor analysis, and main-study validation using confirmatory factor analysis. This article does not report empirical results from that research.

Limitations: what we do not yet know

This article proposes a lens, not a validated conclusion. It does not know, and does not claim to know, which of the four TOE-H dimensions matters most for Mauritian SMEs specifically, whether the human-factor mechanisms described above actually operate the way the broader literature suggests they do in this context, or whether small-island economies behave differently from other SME populations in ways that would need separate evidence to establish. Those are precisely the questions the ongoing doctoral research is designed to test. Until that testing produces analysed results, this piece will remain a conceptual argument, not an evidence report, and it will be revised, with the line between interpretation and evidence redrawn, once that work is further along.