Home/Who It's For/SMEs
You do not have a transformation budget. You still have the problem.
Smaller organisations get the same AI capability problem as large ones, without the learning function, the change programme or the consultancy retainer. The good news is that the honest solution is smaller than the one being sold to enterprises.
What the problem looks like at your size
Everyone is using AI. Nobody agreed how. The person who is best at it is best because they are naturally careful, not because a standard exists. Client-facing work has been through a model at some point and nobody could tell you exactly where. And the thought of formalising any of it competes with actually delivering the work, which wins every time.
The risk is not dramatic. It is that a quietly wrong document reaches a client, and the answer to "how did that happen" is that nothing was ever defined.
The proportionate response
Do not attempt an organisational AI strategy. Take one recurring output that matters — the proposal, the report, the client deliverable, the specification — and design one workflow for it properly. Explicit standards, defined verification, clear allocation of what a human must own, and stop conditions.
That single workflow does more for consistency and defensibility than a policy document, and it is achievable alongside real work. It also becomes the template for the second one.
What exists today
Individual professional pathways — Available. The most practical route for an SME is usually one or two people building the workflow properly and the standard spreading from there.
Organisational benchmarking — by arrangement, and genuinely useful at ten people as well as at a thousand.
Live implementation workshops — run with a small group from one organisation.
Start a conversationWhat does not exist yet
A single packaged SME deployment product at an SME price point. It is on the roadmap and labelled Future: no date, no waitlist promise. The available routes above are the sensible way in meanwhile.
Practical sequence
What we would actually suggest, in order.
01
Profile two people
Whoever produces the most client-facing knowledge work. When the profile opens, you will learn more from two honest baselines than from a survey of everyone.
02
Pick one output
The recurring deliverable where a quality failure would actually cost you something.
03
Build one workflow
Via a workflow pack or a workshop. One workflow, done properly, operated for a quarter.
04
Then spread it
You now have a working standard and evidence of it operating. That is a far easier thing to roll out than an idea.