SME policy pathway: Novice to Champion
The OECD G7 paper maps its descriptive four-stage SME AI taxonomy (Novices → Optimisers → Explorers → Champions) onto a prescriptive policy intervention matrix. Different SMEs at different stages need fundamentally different policy levers. This page captures that matrix and uses it to read across G7 country profiles.
The matrix
| Stage | Defining gap | Dominant enabler need | Representative policy lever |
|---|---|---|---|
| AI Novices | Awareness, basic capability, early-stage finance | Skills + connectivity | Awareness campaigns; advisory; modular guidebooks (e.g. Japan METI’s AI Introduction Guidebook); awareness-stage subsidies (e.g. France IA Booster); SME-targeted start-up loans (UK British Business Bank) |
| AI Optimisers | Workforce preparedness; operational restructuring; risk of scaling investment | Skills + finance + sector adaptation | Co-financed upskilling (UK Flexible AI Upskilling Fund); embedded R&D skills (Canada AI Assist); pre-approved automation subsidies (Japan Catalog-Type); fiscal incentives for digitalisation (Italy Transition 5.0); sectoral grants (UK BridgeAI) |
| AI Explorers | High-performance compute; data quality; long-term R&D capital | AI-enabling inputs + finance | Sovereign compute access (Canada AI Compute Access Fund; UK AIRR; US NAIRR; Germany AI Service Centres); data infrastructure (UK National Data Library; France Current AI); non-dilutive R&D grants (US SBIR/STTR) |
| AI Champions | Frontier-scale ecosystem; world-class talent; integration | Ecosystem + stable business environment | EuroHPC AI Factories (Germany, France, Italy); innovation clusters (France IA-Clusters); compute roadmaps (UK GBP 2bn through 2030); infrastructure pledges (Japan USD 65bn) |
What the matrix is doing
It is doing three things simultaneously:
- Operationalising the oecd-sme-enabler-quartet: the four enablers don’t matter equally at every stage. Skills and connectivity matter most for Novices; compute and data most for Explorers; ecosystems most for Champions.
- Linking diagnosis to prescription: the descriptive taxonomy isn’t theatre — it tells policymakers which instruments to use on which firms.
- Anchoring instrument heterogeneity: a country that spends only on Champion-stage interventions (frontier compute, AI Factories) while neglecting Novice-stage awareness leaves most of its SME base behind. This is the diffusion-vs-frontier debate (see diffusion-vs-innovation) made operational.
Reading G7 profiles against the matrix
Each G7 country emphasises different stages:
- Canada: Strong on Explorer (AI Compute Access Fund) and Optimiser (RAII through Regional Development Agencies, AI Assist).
- France: Strong across the board — IA Booster (Novice/Optimiser), Osez l’IA (Novice broad-base), AI Factory France (Champion), Current AI dataset access (Explorer).
- Germany: Strong on Champion-stage infrastructure (HammerHAI, Jupiter AI Factory) and Optimiser (“Generative AI for SMEs”, KI4KMU); historically heavy on Mittelstand-Digital Innovation Hubs spanning Novice → Optimiser.
- Italy: Optimiser-heavy via Transition 5.0 tax credits and Competence Centres; Champion via IT4LIA AI Factory.
- Japan: Heavy on Novice/Optimiser turnkey support (catalogue subsidies, METI guidebooks) and Champion-stage compute (USD 65bn pledge, Rapidus semiconductors); thin in the middle.
- UK: Comprehensive — BridgeAI (Optimiser-sector), AIRR + Compute Roadmap (Explorer/Champion), SME Digital Adoption Taskforce (broad), Flexible AI Upskilling Fund (Optimiser).
- US: Champion-tilted (NAIRR, NIST, NSF AI Institutes); SME support via SBIR/STTR and SBA AI Resource Hub — relies on private VC for the rest. Most market-driven of the seven.
Open questions
- The matrix is implicit, not stated as a table in the paper; the reconstruction here may overstate its neatness.
- The paper doesn’t grade whether countries match their interventions to their SME distribution. A country with mostly Novices spending mostly on Champion programmes is failing — but the diagnostic remains unanswered.
- Cross-link: this matrix is the buyer-side mirror of taker-shaper-maker. Novices → takers (rent off-the-shelf). Optimisers → takers becoming shapers. Explorers → shapers. Champions → makers.