Most SMEs Don’t Have an AI Problem. They Have a Translation Problem.

The real competitive advantage in the next two years won’t be who has the best AI. It’ll be who can explain, in one short sentence, what they’re actually using it for.

Last month I heard about the managing director of a £14m logistics business slide a £38,000 quote across a boardroom table. It was for an “AI‑powered optimisation platform.” He asked one question: “What will it actually do on Tuesday morning?”

The vendor (an old friend of mine who works for an AI firm who told me what had happened and askd me my thoughts on it) couldn’t answer. Nor could the consultant who’d introduced them.

That small boardroom contained the single biggest obstacle to AI adoption in the UK’s SME economy — and it has nothing to do with models, data, or budget. It’s language.

The translation gap

There is a widening gap between how AI is sold to SMEs and how it actually lives inside a small business.

Vendors describe capabilities (“we use a large language model fine‑tuned on…”). Business owners ask about outcomes (“will this stop my quoting team working Saturdays?”). Each side believes the other is being unreasonable. Both walk away frustrated, and the pilot either stalls or gets signed off out of fear of being left behind — the two most expensive ways to buy technology.

This is the translation gap, and it is doing quiet, expensive damage across the mid‑market. SME AI adoption is running well below intention, and the usual explanation — “they’re nervous, they need more education” — is half right and entirely the wrong frame. SME owners aren’t slow to adopt AI. They’re slow to adopt AI they can’t describe. And they’re right to be.

Why the gap exists

Three forces keep the translation gap wide.

First, vendor language is built for other vendors. Most AI marketing is optimised to impress a buying committee of technical reviewers, not a single MD whose day is already full. “Enterprise‑grade retrieval‑augmented generation” means nothing to someone who just wants their invoices paid faster. The more sophisticated the product, the more it talks past its smallest customers.

Second, nobody wants to look behind. AI is the first technology in a generation where senior leaders feel pressure to nod along even when they don’t understand. That creates a very specific failure mode: projects approved on vibes, measured on vanity, and quietly shelved six months later.

Third, we keep treating AI like an IT project. It isn’t. IT projects replace an existing process with a faster version of the same process. AI projects change what the process is. That distinction sounds abstract, until you realise the project plan you used for the Microsoft 365 migration won’t survive contact with a system that generates new content each time it runs.

The three-sentence test

Here is the single most useful tool I can give any SME owner weighing an AI decision. Before you sign anything — a contract, a pilot, an internal mandate — insist that the initiative can be explained in three short sentences that your receptionist would understand.

1.     What problem are we solving? Name it plainly. “Quotes take four days and we lose deals because of it.” Not “we are pursuing operational excellence through intelligent automation.”

2.    Who benefits, and how? Named people, named tasks, measured in hours or pounds. “Sarah and the two estimators get back Friday afternoons. We expect to win three more quotes a month at an average margin of £4,200.”

3.    How will we know it worked? A number, a date, and a person who owns it. “By 30 June, our average quote turnaround is under 24 hours, or we stop.”

If you cannot produce those three sentences, you do not have an AI project. You have an AI purchase. They are not the same thing, and only one of them creates value.

This test sounds almost embarrassingly simple. That is the point. It is simple enough to survive the five‑week distortion field between a good demo and a signed contract. It is simple enough for a chair, a non‑exec, a finance director, and a warehouse manager to all agree on in the same meeting. Simplicity, in AI adoption, is not a starting condition. It is the deliverable.

What AI Adoption looks like in practice

A client of ours — a family‑run manufacturing business in the Midlands, 70 staff, no in‑house IT — ran this test on an AI proposal we didn’t write.

The vendor’s pitch was “intelligent production planning.” Impressive demo, forty‑slide deck, £22k a year.

Their three sentences, once we helped them draft them honestly, read: “We spend about six hours a week rebuilding the production plan when orders shift. We want our planner, Raj, to do that in under an hour. By the end of Q3, we want to ship two more orders a week on time.”

The proposal, it turned out, did none of those things. It optimised a different problem beautifully — one they didn’t have. They passed, re‑scoped the brief, and ended up buying a far cheaper tool that actually moved the metrics they cared about. The translation exercise saved them roughly £60k over two years and, more importantly, saved Raj’s Friday evenings.

Three things to do on Monday to get AI Adoption Right

If you lead an SME and you want to close your own translation gap this week:

Write your three sentences for the AI initiative currently highest on your desk. If you can’t, your team can’t either — and the project is already in trouble.

Ban the word “AI” from your next internal conversation about it. Force the discussion back to the job you are trying to get done. You will be surprised how much clarity appears when the acronym leaves the room.

Appoint a translator, as a champion. The person who succeeds with AI in your business won’t be the most enthusiastic. It’ll be the one who can move fluently between “what the tool does” and “what our Tuesday looks like” and champions it.

Clarity is the strategy

The companies that will pull ahead over the next 24 months are not the ones with the cleverest models or the biggest budgets. They are the ones whose managing directors can walk into any room in their business — the warehouse, the showroom, the finance office, the boardroom — and describe, in plain English, exactly what they are asking AI to do, for whom, and by when.

That is not a technology capability. It is a leadership one. And, as it turns out, it is the single highest‑leverage thing any SME owner can invest in this year.

AI is not hard. Deciding what to use it for is hard. We keep mistaking one for the other, and it is costing the UK’s best small businesses real money.

The translation gap is closable. But only by the people running the business. Nobody is coming to do it for you — and, quietly, that is very good news.