We are based in Llanelli. That is a deliberate fact rather than an accident of where somebody happened to live, and it changes what we see.
Most published advice about adopting AI is written from and for a particular kind of business: venture-funded, urban, forty people, already running a modern stack, with a head of operations whose job is finding efficiencies. That advice is not wrong. It is just aimed at a company that does not exist in most of Wales.
Here is what is actually different, what is not, and what tends to work.
What is genuinely different
Teams are smaller than the advice assumes
The standard playbook says: appoint an owner, define success criteria, run a pilot, review monthly. Good advice. It assumes there is someone whose diary can absorb it.
In a nine-person firm the owner of any new initiative is the person who also does quoting, invoicing and half the delivery. That is not a resourcing failure to be fixed (it is the shape of the business. What follows is that anything requiring sustained attention from that person will lose to the work that pays this month, every time.
What works instead: pick things that run without supervision once built. Answering, acknowledgement, reminders, follow-up sequences. Avoid anything that needs a human to check its output daily, however clever it is, because that check will not survive a busy fortnight.
Distance is real, and it cuts both ways
A workshop day in Carmarthen or Aberystwyth is a different proposition from one in Cardiff. That is why most consulting days run remotely and on-site days are booked for the formats that genuinely need a room.
But the more interesting effect is on the businesses themselves. A firm covering West Wales is already used to travel being expensive, which makes them unusually receptive to anything that reduces wasted journeys. No-show reduction and better job intake) both dull, both on our trades list (land harder here than they do in a dense city, because a wasted visit costs an hour of driving rather than fifteen minutes.
Seasonality is sharper
Tourism, agriculture and construction dominate more of the local economy than national averages suggest, and they are seasonal in a way that professional services are not.
This changes the roadmap sequencing rather than its content. Building through the quiet season and going live before the busy one is the difference between a project that gets adopted and one that gets abandoned in week two because nobody has time to learn it. Any plan that puts a go-live in the middle of your peak has been written by someone who has not asked.
Welsh language
If you serve the public in Wales, bilingual capability is a service question and sometimes a procurement requirement, particularly with public sector work.
Modern models handle Welsh considerably better than they did even two years ago, but "handles Welsh" and "handles Welsh well enough to represent your business" are different bars, and the gap is wider for spoken output than written. Test it properly with a fluent speaker before it faces a customer. Do not take a vendor's word for it.
What is not different at all
This is the shorter list, and the more important one.
The bottlenecks are identical. Unanswered calls, follow-up that depends on who is free, data carried between systems by a human, and a process that lives in one person's head. We find the same four in Llanelli, Cardiff, Bristol and London. The audit finds them everywhere.
The failure modes are identical. Pilots die in month three for the same four reasons regardless of postcode.
The technology is identical. There is no Welsh discount and no Welsh penalty on the tools. A business in Llanelli has access to exactly what a business in Shoreditch has, at exactly the same price. Any suggestion otherwise is either sales or fatalism.
That third point deserves emphasis, because the fatalism is real and it is expensive. We have sat with capable businesses who assumed this sort of work was for somewhere else. It is not. The barrier is knowing which thing to do first, and that is a conversation, not a location.
What tends to work here
Start with the phone. Welsh SMEs skew towards trades, home services, hospitality, agriculture-adjacent businesses and local professional services. All of those are phone-first. The missed-call report is usually the highest-value thing anyone has looked at that quarter.
Fix the dull thing before the interesting one. A business that automates on top of an undefined process gets the undefined process at speed. Where a team is small, the cost of that mistake is proportionally larger, because there is no slack to absorb it.
Build during the quiet season. Genuinely. Sequence the roadmap around the trading calendar, not the financial year.
Keep the tool count down. Seven subscriptions held together by integrations is a maintenance load, and in a nine-person firm there is nobody to carry it. Fewer, better-configured tools beat a broader stack.
Train the people who will use it, not just the owner. Adoption stalls more often from confusion than from bad software. That is true everywhere and it bites harder when there are four people rather than forty, because one confused person is 25% of your team.
On grants and support
There is more public support available to Welsh businesses for this kind of work than most owners realise, through Business Wales and various regional programmes, and the picture changes often enough that specifics go out of date quickly.
Two honest notes. First: check what is currently available before you commit budget, because it genuinely moves the arithmetic on a first project. Second: do not let the availability of funding decide what you build. Funded work aimed at the wrong bottleneck is still wasted, and it is worse than wasted if it uses up an allowance you could have spent on the right thing.
Work out the right thing first. Then find out whether it can be funded.
Where we work
The office is in Llanelli and most days run remotely, which means the coverage is genuinely UK-wide rather than a claim on a website. On-site is easiest across West and South Wales) Swansea, Carmarthen, Llanelli, Cardiff, Newport: then Bristol and the South West. Further afield is booked as workshop days rather than pretended to be a local office.
Same retainers, same method, same people everywhere. There is no regional price list, and the prices are published.
If you run a business in Wales and want a straight conversation about where your enquiries and hours are leaking, the AI Opportunity Assessment is free and takes about forty-five minutes. It is a diagnosis, not a pitch, and a reasonable proportion end with us saying you do not need us yet.