A strategy paper for SME leaders adopting AI
The AI4SME Resource Guide to 2026
Who this is for: the owner, managing director, or business leader of a small or medium enterprise who knows AI matters but is unsure where to start, what to believe, and how to avoid an expensive false start.
Current as of August 2026. Built on AIRI v3.1 and nine years of AI programme experience across more than one thousand organisations — from "why bother" through to a deployed AI solution and a stronger team.
Get the guide
The download is being finalised. Meanwhile, email contact@ai4sme.org with “guide” in the subject and it will be sent to you.
Get the guide
The download is being finalised. Meanwhile, email contact@ai4sme.org with “guide” in the subject and it will be sent to you.
The playbook, the evidence, and the eight questions
The playbook
Assess, then build — the five moves, matched to your maturity level, that take a real bottleneck to a deployed solution.
The evidence
3,000+ organisations assessed, 300+ AI projects delivered — and what the deployments actually returned, counted as plainly as the failures.
The eight questions
Ask any provider these before signing. A provider that cannot answer them is selling you a model.
What's inside
- 1
The SME AI imperative
Why this moment is different for SMEs, stated honestly.
- 2
Assess your readiness with AIRI v3.1
One framework, two instruments, five pillars, five levels — and your target.
- 3
The SME AI adoption playbook
Match the engagement to your maturity; the five moves.
- 4
Engagement pathways
Build with help, or build it yourselves — and win the apprentice.
- 5
What success looks like
Real deployments, flagship and SME-sized.
- 6
Choosing and vetting an AI solution provider
The eight questions to ask before you sign.
- 7
Measuring your progress
Metrics that reflect readiness and outcomes, not vanity activity.
- 8
Recommendations for the SME leader
The whole guide in eight moves.
- A/B
Glossary · Sources and further reading
The terms defined by the job they do, and where the numbers come from.
Read this before downloading
Most SMEs know they should be doing something with AI. Most are not sure what, and the market is crowded with vendors promising transformation and delivering pilots that never reach production. The result is a familiar pattern: a model that works in a demo, a project that stalls, and a leadership team that concludes "AI is not for us."
That conclusion is almost always wrong. The reasons AI projects fail rarely involve the technology. They involve readiness. The fix is to assess where you stand before you build, then engage on a project that is real, bounded, and built with people who will still be there when the consultants leave.
The single most important message in the guide: for most SMEs, the destination is AI Ready (AIRI Level 2) — and that is a target, not a consolation.
Your target is Level 2
Unlike assessments that push everyone toward "expert," AIRI is explicit: AI Ready (Level 2) — simple department-level workflows, often built by your own domain experts — is where the gains are for most organisations. Very few SMEs should aim for Level 3, which means an in-house AI team most will never have.
The five levels explainedThe eight questions to ask a prospective provider
Free here, in the guide with the full vetting context. A provider that answers them in detail is selling you a deployed capability.
1. Delivery track record
How many AI solutions have you delivered into production (not demos)? Can I speak to those customers?
2. Production bar
Will the deliverable run in my systems, behind an API, or will I receive a notebook? Insist on the former.
3. Data work
Who handles data acquisition, cleaning, and ongoing maintenance? How is data quality sustained after handover?
4. Monitoring and retraining
How will you monitor the live model, and what triggers a retrain? A model that is not monitored will drift.
5. The team
Who specifically will work on my project, and what is their shipping experience (as distinct from their research background)?
6. Intellectual property
Who owns the IP, the data, and the model weights when the engagement ends? You should.
7. Ethics and governance
How do you test for bias, handle sensitive data, and manage the risks of AI agents acting without a human checking?
8. Total cost of ownership
What does running this cost me in year two and year three, not just to build?
Questions about the guide
- Is it really free?
- Yes — free, and licensed CC BY 4.0: share it with attribution. The email gate exists so the guide can reach the people it was written for.
- How long does it take to read?
- Eighteen pages, written to be read in one sitting and argued with your leadership team in the next meeting.
- Who is it for?
- The owner, managing director, or business leader of an SME who knows AI matters but is unsure where to start, what to believe, and how to avoid an expensive false start.
Assess first. Then build.
The guide walks the whole path — the free assessment is its first step.
Take the free assessment