An AI Strategy for a Non-Tech Business
Share
An AI strategy that runs to forty slides and changes nobody's Monday morning is not a strategy. It is a document.
Direct answer
An AI strategy for a normal business is three decisions on one page. Which tasks you will target, what staff must never do, and who owns the outcome. Everything else is implementation detail dressed up as strategy.
Based on the AI and Automation Practitioner standard (ST1512), a Level 4 apprenticeship for people already in work. Funding rules described here are the England position and are set by the funding body.
Potentially free for you
The learner never pays. Employers potentially pay nothing either, and where the apprentice is aged 16 to 24 the government funds the whole cost even if the employer does not pay the levy.
Closing any gap. A levy transfer from another employer covers 100% of the cost. Unused transfer allowance is not carried forward at the end of the tax year, and it must be agreed before the apprentice starts.
From 1 August 2026 the government stopped adding the 10% top-up to new funds entering levy accounts, so there is less in the system than there was.
Part of our guide to AI for UK Business: Where It Pays and Where It Does Not.
Decision one: which tasks
Name three to five specific recurring tasks. Not functions, not departments, tasks. First-line replies, quote assembly, meeting actions.
Pick them because they are text-heavy, frequent and cheap to check. Not because they are strategic.
Write the current time cost next to each. Without that number you cannot tell later whether anything happened.
Anything not on this list is out of scope for the year. That constraint is what makes it a strategy rather than a wish list.
Decision two: the boundaries
What must never be pasted into a tool. Customer records, employee data, contract terms, anything under an NDA. Name them specifically.
Which tools are approved. A short list is easier to defend, easier to support and easier to train on than an open one.
What output must be checked by a person before it reaches a customer, and who that person is.
These three lines prevent most of what goes wrong. They fit on half a page and they need no legal review.
Decision three: who owns it
Name a person, not a committee. Committees produce documents and individuals produce change.
Give them the three to five tasks and a date to report the measured result. Six months is about right.
Give them time. An owner with no protected hours is a name on a slide, and everyone in the organisation can tell.
Make the report a number rather than a narrative. Minutes saved on each named task, with the rewrite rate alongside.
There is more on this in our guide to AI for UK businesses.
What to leave out
Predictions about the industry. They date badly, nobody acts on them, and they are the reason these documents get ignored.
Tool comparisons. Tools change every few months and a strategy tied to a product is obsolete by spring.
Headcount implications, unless you have actually decided something. Speculating about jobs in a strategy document damages trust for no benefit.
Anything you could not explain to a new starter in two minutes. Length is not depth.
How to know it is working
Somebody's week is visibly different within a month. If nobody's calendar changed, nothing was implemented.
The owner can quote a number without preparing. Numbers that need preparing are constructed rather than tracked.
People outside the project start asking for the tool. Pull is a better signal than any adoption statistic.
If none of those are true at six months, the strategy did not fail. It was never implemented, which is a different problem with a different fix.
What UK employers should include
A line on data protection. Naming what must never be pasted in satisfies most of the practical UK GDPR question in one sentence.
A line on funding. If you pay the apprenticeship levy you have money sitting unspent that can only be used on training, and it expires.
A line on your sector regulator. UK regulation is regulator-led rather than statute-led, so the guidance that applies to you is specific to your sector.
None of those need a consultant. They need someone in the business to decide and write them down.
The capability question underneath it
Every one of these decisions assumes someone in the business understands how these tools fail. Frequently nobody does.
Buying more tools does not fix that, and consultants leave with the understanding they brought.
A funded apprenticeship builds it in one of your existing staff over 14 months, while they keep doing the job.
For a levy-paying employer it comes from money already contributed. For others the government funds 95 per cent, and a levy transfer from a larger employer can cover the rest.
Key takeaways
- Three decisions on one page beats forty slides.
- Name three to five specific tasks with their current time cost.
- Name what must never be pasted in, and who checks output.
- Give one person ownership, protected time and a numeric report date.
- Leave out predictions, tool comparisons and headcount speculation.
Frequently asked questions
What should an AI strategy include?
Which specific tasks you will target, what staff must never do, and who owns the outcome. Three decisions on one page.
Do small businesses need an AI strategy?
They need the three decisions, which take an afternoon. They do not need a consultancy framework or a forty-slide document.
How many tasks should we target?
Three to five specific recurring tasks, with the current time cost written next to each. Anything else is out of scope for the year.
Who should own AI in a business?
A named person with protected time and a date to report a measured result. Committees produce documents rather than change.
What should we leave out of the strategy?
Industry predictions, tool comparisons, and headcount speculation. All three date badly and none of them change anyone's week.
How do we measure whether it worked?
Minutes saved on each named task with the rewrite rate alongside. A narrative report tells you how people feel rather than what happened.
Do we need consultants?
Rarely. The three decisions require knowledge of your own business, which consultants do not have and cannot leave behind.
What if nobody understands the technology?
That is the actual gap, and more tools will not close it. Training one existing person properly is the cheaper and more durable answer.
The funded route, in figures
Figures are the England position. Levy-paying employers usually draw the whole cost from the levy they already contribute, so they pay nothing extra. Employers below the £3 million pay bill threshold pay nothing for an apprentice aged 16 to 24. For an older apprentice the government funds 95 per cent, and a levy transfer from a larger employer can cover the rest. Eligibility and funding rules are set by the funding body and can change.
Where to go next
If the missing piece is someone who genuinely understands where these tools fail, the funded route builds that person inside your business.
We help UK employers and employees access government-funded AI training. Training is delivered by third-party providers. We do not guarantee a place, a qualification, a job or any particular outcome. Eligibility and funding rules are set by the funding body and can change. Always check the current position before you commit. See our disclaimer.