OECD SME enabler quartet
The OECD G7 discussion paper on SME AI adoption identifies four enablers that must be in place for AI to diffuse into the SME segment of an economy: (i) connectivity, (ii) AI-enabling inputs (data, algorithms, compute), (iii) skills, and (iv) finance. These are pulled forward from the OECD AI Principles and the OECD SME and Entrepreneurship Policy recommendation, and they structure both the diagnostic and the policy half of the paper.
The four enablers
- Connectivity — high-speed broadband, fibre, 5G, FWA. A pre-condition for any downstream digital uptake. The OECD frames this through both firm-size and urban-rural divides (see sme-connectivity-divide).
- AI-enabling inputs — the “AI production function” (OECD, 2023): data, algorithms, compute. Relevant for SMEs that want to train or fine-tune, but also for those choosing whether to deploy custom vs off-the-shelf. SMEs predominantly access compute via hyperscaler clouds (AWS, Azure, GCP); short-term GPU contracts and tiered pricing are flagged as start-up pain points (Canada survey, 2024).
- Skills — both ICT/technical and complementary (critical thinking, management, creativity). Consistently the top-reported barrier (see ai-skill-shortage-as-diffusion-bottleneck).
- Finance — credit access, equity, fintech, alternative finance. Tightening credit conditions since 2022 have starved long-term technology investment in SMEs (see sme-ai-finance-gap).
Relationship to the WEF/Accenture six-factor
The OECD quartet is a narrower, SME-specific, infrastructure-and-input-focused cousin of foundational-enablers (WEF/Accenture’s ecosystem, trust, self-governance, talent, cybersecurity, digital core).
| OECD quartet | WEF six-factor | Notes |
|---|---|---|
| Connectivity | (subsumed in digital core) | OECD treats it as a distinct precondition |
| AI-enabling inputs | (subsumed in digital core / ecosystem) | OECD breaks data/algos/compute apart |
| Skills | Talent | Direct overlap |
| Finance | — | OECD-only; WEF assumes enterprise-scale balance sheets |
| — | Trust, Self-governance | WEF-only; OECD relegates to “light-touch regulation” |
| — | Cybersecurity | OECD touches in passing |
The WEF frame is enterprise-side and governance-heavy; the OECD frame is SME-side and input-heavy. Together they cover the full demand-side enabler space. For the white paper, the OECD quartet should be used when discussing diffusion barriers in the long tail of the firm distribution; the WEF six-factor when discussing what stops a large enterprise from reaching ROI on a pilot.
Heterogeneity caveat
The paper is explicit that “one size does not fit all”: the relative weight of each enabler depends on a firm’s digital maturity and on the complexity of the AI use case it pursues. Turnkey generative AI tools shift the burden toward skills and finance; custom solutions push back on inputs and compute. This is the bridge into the four-stage sme-policy-pathway-novice-to-champion taxonomy.
Connection to working hypotheses
The quartet is a demand-side, buyer-side complement to the supply-side H1_L0-L7-ladder. The supply side asks where value lands in the stack; the OECD demand side asks which firms can actually reach far enough up the stack to capture value at all. Where any enabler is missing, that firm cannot meaningfully participate at the application layer — thinning out the bottom-right of the H2_u-curve-of-value.
Also directly supports diffusion-vs-innovation: the entire enabler frame is a diffusion playbook.
See also
- theoretical-vs-observed-capability-gap — worker/occupation-level version of the same binding-constraint logic; the firm-level enabler quartet has a worker-level mirror.